Mythos is here, it’s time to start tokenmaxxing
Read full transcript 24 segments
-
You might have noticed I haven't been You might have noticed I haven't been posting as many videos the last few posting as many videos the last few posting as many videos the last few days, and there's a good reason for it. days, and there's a good reason for it. days, and there's a good reason for it. It is indeed the Fable release. It's It is indeed the Fable release. It's It is indeed the Fable release. It's such a powerful model, and I've been such a powerful model, and I've been such a powerful model, and I've been pushing it to its absolute limits to the pushing it to its absolute limits to the pushing it to its absolute limits to the point where it's affecting my sleep, my point where it's affecting my sleep, my point where it's affecting my sleep, my day-to-day, and of course, the content day-to-day, and of course, the content day-to-day, and of course, the content I'm putting out. And I don't want to do I'm putting out. And I don't want to do I'm putting out. And I don't want to do a normal video where I just show some of a normal video where I just show some of a normal video where I just show some of the cool things I've been doing with it, the cool things I've been doing with it, the cool things I've been doing with it, because this release is different. As because this release is different. As because this release is different. As you'll probably know by now, Fable, aka you'll probably know by now, Fable, aka you'll probably know by now, Fable, aka Mythos with some safeguards, is Mythos with some safeguards, is Mythos with some safeguards, is available on the Pro and Max tier available on the Pro and Max tier available on the Pro and Max tier subscriptions on Claude Code, but only subscriptions on Claude Code, but only subscriptions on Claude Code, but only until June 22nd, because as of June until June 22nd, because as of June until June 22nd, because as of June 23rd, it will be removed from those 23rd, it will be removed from those 23rd, it will be removed from those plans. The model seems to be just too plans. The model seems to be just too plans. The model seems to be just too expensive and too compute-heavy for them expensive and too compute-heavy for them expensive and too compute-heavy for them to run in the subscription tier that is to run in the subscription tier that is to run in the subscription tier that is as subsidized as it is. And those as subsidized as it is. And those as subsidized as it is. And those subsidies are crazy. We're talking like subsidies are crazy. We're talking like subsidies are crazy. We're talking like $8,000 a month. These plans can get you $8,000 a month. These plans can get you $8,000 a month. These plans can get you up to $8,000 a month of inference for up to $8,000 a month of inference for up to $8,000 a month of inference for just $200. And I'm not talking out of my just $200. And I'm not talking out of my just $200. And I'm not talking out of my ass here, I'm speaking from experience. ass here, I'm speaking from experience. ass here, I'm speaking from experience. I've spent the last few days going hard I've spent the last few days going hard I've spent the last few days going hard on my subscriptions for Claude Code, as on my subscriptions for Claude Code, as on my subscriptions for Claude Code, as well as Codex to an extent, and I've well as Codex to an extent, and I've well as Codex to an extent, and I've managed to do $4,358 managed to do $4,358 managed to do $4,358 of inference in the last 10 days alone.
-
of inference in the last 10 days alone. of inference in the last 10 days alone. There's a catch though, that's just on There's a catch though, that's just on There's a catch though, that's just on this laptop. I also have a Mac Mini I've this laptop. I also have a Mac Mini I've this laptop. I also have a Mac Mini I've been doing a lot of work on too, and been doing a lot of work on too, and been doing a lot of work on too, and this one's up to $1,112 this one's up to $1,112 this one's up to $1,112 of additional inference on top of all of of additional inference on top of all of of additional inference on top of all of this. That's a shitload of tokens, and I this. That's a shitload of tokens, and I this. That's a shitload of tokens, and I want to show how I've been using them to want to show how I've been using them to want to show how I've been using them to get real work done. And I want to be get real work done. And I want to be get real work done. And I want to be clear, not all of this is super useful. clear, not all of this is super useful. clear, not all of this is super useful. This is not a video about maximizing how This is not a video about maximizing how This is not a video about maximizing how much you get per token or per dollar. much you get per token or per dollar. much you get per token or per dollar. It's kind of the opposite. It's how you It's kind of the opposite. It's how you It's kind of the opposite. It's how you can take advantage of these really can take advantage of these really can take advantage of these really generous subsidized subscription plans generous subsidized subscription plans generous subsidized subscription plans in the 10 days we have left with people in the 10 days we have left with people in the 10 days we have left with people on them, which is why I'm rushing this on them, which is why I'm rushing this on them, which is why I'm rushing this video out, of course, so that you can video out, of course, so that you can video out, of course, so that you can get the most possible value and see and get the most possible value and see and get the most possible value and see and taste what the future can kind of look taste what the future can kind of look taste what the future can kind of look like with models this powerful, assuming like with models this powerful, assuming like with models this powerful, assuming cost is no issue. Obviously, this is not cost is no issue. Obviously, this is not cost is no issue. Obviously, this is not realistic long-term. Nobody should be realistic long-term. Nobody should be realistic long-term. Nobody should be spending 10 grand a month personally on spending 10 grand a month personally on spending 10 grand a month personally on inference for day-to-day work, but inference for day-to-day work, but inference for day-to-day work, but spending two to $400 and getting that spending two to $400 and getting that spending two to $400 and getting that much inference, that is kind of much inference, that is kind of much inference, that is kind of compelling. I've never burned so many compelling. I've never burned so many compelling. I've never burned so many tokens in my life. In fact, I'm pretty tokens in my life. In fact, I'm pretty tokens in my life. In fact, I'm pretty sure I I more in the last 10 days than I sure I I more in the last 10 days than I sure I I more in the last 10 days than I did in the rest of my life prior to that did in the rest of my life prior to that did in the rest of my life prior to that and I've learned a ton since. The good, and I've learned a ton since. The good, and I've learned a ton since. The good, the bad, the ugly and more. From loops the bad, the ugly and more. From loops the bad, the ugly and more. From loops to open claw and Hermes agent to crazy to open claw and Hermes agent to crazy to open claw and Hermes agent to crazy workflows that will automate value being workflows that will automate value being workflows that will automate value being poured out of your code bases and more.
-
poured out of your code bases and more. poured out of your code bases and more. There's a lot to dig into here and a lot There's a lot to dig into here and a lot There's a lot to dig into here and a lot of cool opportunities to maximize your of cool opportunities to maximize your of cool opportunities to maximize your usage of these tools and models. But usage of these tools and models. But usage of these tools and models. But before I can dive into all that, I before I can dive into all that, I before I can dive into all that, I highly recommend maximizing on today's highly recommend maximizing on today's highly recommend maximizing on today's sponsor. AI has gotten great at coding sponsor. AI has gotten great at coding sponsor. AI has gotten great at coding but deploying, not so much. As great as but deploying, not so much. As great as but deploying, not so much. As great as things like AWS are, good luck trying to things like AWS are, good luck trying to things like AWS are, good luck trying to configure things properly using an configure things properly using an configure things properly using an agent, especially when you're trying to agent, especially when you're trying to agent, especially when you're trying to do multiple things, have preview do multiple things, have preview do multiple things, have preview environments, spin your team up as well. environments, spin your team up as well. environments, spin your team up as well. It's just not a great story and that's It's just not a great story and that's It's just not a great story and that's why Render is so, so cool. These guys why Render is so, so cool. These guys why Render is so, so cool. These guys built an enterprise-ready cloud that is built an enterprise-ready cloud that is built an enterprise-ready cloud that is also agent-ready. While every other also agent-ready. While every other also agent-ready. While every other cloud is struggling to keep up with cloud is struggling to keep up with cloud is struggling to keep up with what's going on, they're making nice what's going on, they're making nice what's going on, they're making nice changes like dropping the per-seat changes like dropping the per-seat changes like dropping the per-seat pricing, instead just charging you for pricing, instead just charging you for pricing, instead just charging you for your infrastructure. So it doesn't your infrastructure. So it doesn't your infrastructure. So it doesn't matter how many people or agents you matter how many people or agents you matter how many people or agents you have on your team, all that matters is have on your team, all that matters is have on your team, all that matters is how much you're shipping. You can use how much you're shipping. You can use how much you're shipping. You can use Render to host anything. Web services, Render to host anything. Web services, Render to host anything. Web services, databases, cron jobs, workflows, static databases, cron jobs, workflows, static databases, cron jobs, workflows, static sites, CDN content, KVs, everything is sites, CDN content, KVs, everything is sites, CDN content, KVs, everything is supported by these guys and they have supported by these guys and they have supported by these guys and they have integrations for everything you'd ever integrations for everything you'd ever integrations for everything you'd ever want. And as I mentioned before, these want. And as I mentioned before, these want. And as I mentioned before, these guys are actually enterprise-ready. They guys are actually enterprise-ready. They guys are actually enterprise-ready. They have private networks built in so your have private networks built in so your have private networks built in so your systems don't have to communicate over systems don't have to communicate over systems don't have to communicate over the internet. This is huge for building the internet. This is huge for building the internet. This is huge for building real systems. And you can configure this real systems. And you can configure this real systems. And you can configure this all with blueprints, simple YAML files all with blueprints, simple YAML files all with blueprints, simple YAML files that describe your entire that describe your entire that describe your entire infrastructure. Imagine if Terraform infrastructure. Imagine if Terraform infrastructure. Imagine if Terraform went way deeper and further, that's what went way deeper and further, that's what went way deeper and further, that's what you get with their blueprints. That you get with their blueprints. That you get with their blueprints. That said, if you want Terraform, they said, if you want Terraform, they said, if you want Terraform, they support that too, don't worry. And if support that too, don't worry. And if support that too, don't worry. And if you're building heavily distributed work you're building heavily distributed work you're building heavily distributed work with lots of pieces with high failure with lots of pieces with high failure with lots of pieces with high failure rates, Render workflows are going to rates, Render workflows are going to rates, Render workflows are going to save your butt. It's a simple system for save your butt. It's a simple system for save your butt. It's a simple system for queuing and scheduling your work that queuing and scheduling your work that queuing and scheduling your work that has a great SDK for TypeScript and for has a great SDK for TypeScript and for has a great SDK for TypeScript and for Python, making it easy to make Python, making it easy to make Python, making it easy to make resilient, durable workloads on the same resilient, durable workloads on the same resilient, durable workloads on the same cloud that hosts your database, your CDN cloud that hosts your database, your CDN cloud that hosts your database, your CDN and more. If you're not convinced yet, and more. If you're not convinced yet, and more. If you're not convinced yet, would $50 convince you cuz they'll give would $50 convince you cuz they'll give would $50 convince you cuz they'll give you 50 bucks of credit if you use code
-
you 50 bucks of credit if you use code you 50 bucks of credit if you use code render-theo. Try it now at render-theo. Try it now at render-theo. Try it now at solid.link/render. solid.link/render. solid.link/render. Before we talk about burning all of your Before we talk about burning all of your Before we talk about burning all of your tokens, I want to talk about getting as tokens, I want to talk about getting as tokens, I want to talk about getting as many as possible, as in maximizing your many as possible, as in maximizing your many as possible, as in maximizing your use of the limits that you have. As I use of the limits that you have. As I use of the limits that you have. As I mentioned in my last video, I'm actually mentioned in my last video, I'm actually mentioned in my last video, I'm actually dual wielding right now. I have two dual wielding right now. I have two dual wielding right now. I have two Claude code $200 plans that I rotate Claude code $200 plans that I rotate Claude code $200 plans that I rotate between whenever I max one out, and I between whenever I max one out, and I between whenever I max one out, and I found a lot of ways to gamify that that found a lot of ways to gamify that that found a lot of ways to gamify that that we'll talk about in a second. But first, we'll talk about in a second. But first, we'll talk about in a second. But first, I want to talk a little bit about how I want to talk a little bit about how I want to talk a little bit about how these rate limits work. This is my these rate limits work. This is my these rate limits work. This is my second account, which I haven't had to second account, which I haven't had to second account, which I haven't had to use today, which means that my 5-hour use today, which means that my 5-hour use today, which means that my 5-hour session hasn't started. This is the session hasn't started. This is the session hasn't started. This is the first important thing to start hacking first important thing to start hacking first important thing to start hacking around. You want this timer resetting as around. You want this timer resetting as around. You want this timer resetting as often as possible. Realistically often as possible. Realistically often as possible. Realistically speaking, I can hit the 5-hour limits speaking, I can hit the 5-hour limits speaking, I can hit the 5-hour limits now in about an hour. So, if I don't now in about an hour. So, if I don't now in about an hour. So, if I don't have this timer ticking, and I start have this timer ticking, and I start have this timer ticking, and I start this work session that will kill my this work session that will kill my this work session that will kill my limit in an hour, then I then have to limit in an hour, then I then have to limit in an hour, then I then have to wait 4 hours for it to reset because the wait 4 hours for it to reset because the wait 4 hours for it to reset because the timer doesn't start until you send a timer doesn't start until you send a timer doesn't start until you send a message. Thankfully, this is easy to message. Thankfully, this is easy to message. Thankfully, this is easy to work around. This is all it takes. I'm work around. This is all it takes. I'm work around. This is all it takes. I'm just in the Claude.ai site. I'm going to just in the Claude.ai site. I'm going to just in the Claude.ai site. I'm going to say, "Hi." I'm going to send. I'm going say, "Hi." I'm going to send. I'm going say, "Hi." I'm going to send. I'm going to stop before it does any real work.
-
to stop before it does any real work. to stop before it does any real work. And now the timer is ticking. It will And now the timer is ticking. It will And now the timer is ticking. It will now reset in 4 hours and 50 minutes even now reset in 4 hours and 50 minutes even now reset in 4 hours and 50 minutes even though I haven't dinged any of my usage. though I haven't dinged any of my usage. though I haven't dinged any of my usage. Okay, there's a tiny, tiny bit was Okay, there's a tiny, tiny bit was Okay, there's a tiny, tiny bit was probably used, but basically nothing. probably used, but basically nothing. probably used, but basically nothing. Doing this gets you reset as soon as Doing this gets you reset as soon as Doing this gets you reset as soon as possible. So, now if in 4 [snorts] possible. So, now if in 4 [snorts] possible. So, now if in 4 [snorts] hours, let's say 4 and 1/2 hours, I hours, let's say 4 and 1/2 hours, I hours, let's say 4 and 1/2 hours, I start working, I don't have to worry start working, I don't have to worry start working, I don't have to worry about hitting my limit cuz it's going to about hitting my limit cuz it's going to about hitting my limit cuz it's going to get reset in 30 minutes. That's only one get reset in 30 minutes. That's only one get reset in 30 minutes. That's only one of the limits you have to worry about of the limits you have to worry about of the limits you have to worry about though because there are two. You have though because there are two. You have though because there are two. You have the current session limit, and the current session limit, and the current session limit, and separately, you have the all models separately, you have the all models separately, you have the all models weekly limit, which resets once a week. weekly limit, which resets once a week. weekly limit, which resets once a week. Similar to the 5-hour window, this one Similar to the 5-hour window, this one Similar to the 5-hour window, this one starts counting down after you send your starts counting down after you send your starts counting down after you send your first message. Again, incentivizing you first message. Again, incentivizing you first message. Again, incentivizing you to make sure you go hop somewhere and to make sure you go hop somewhere and to make sure you go hop somewhere and trigger something that gets these limits trigger something that gets these limits trigger something that gets these limits burning. I want to automate this though, burning. I want to automate this though, burning. I want to automate this though, which you'll notice is a theme for a lot which you'll notice is a theme for a lot which you'll notice is a theme for a lot of the things I want to show off here. I of the things I want to show off here. I of the things I want to show off here. I could keep going to the website. I could could keep going to the website. I could could keep going to the website. I could even have a browser use agent go there even have a browser use agent go there even have a browser use agent go there for me, but at this point in time, for me, but at this point in time, for me, but at this point in time, Claude-p still counts towards your rate Claude-p still counts towards your rate Claude-p still counts towards your rate limits. That will change in the near limits. That will change in the near limits. That will change in the near future, which means it probably won't future, which means it probably won't future, which means it probably won't work for that long, but at the very work for that long, but at the very work for that long, but at the very least, the strategy will work now. So, least, the strategy will work now. So, least, the strategy will work now. So, I'm going to set something up. This is I'm going to set something up. This is I'm going to set something up. This is not meant to be like, "Oh, look look how not meant to be like, "Oh, look look how not meant to be like, "Oh, look look how smart I am." It's meant to try and get smart I am." It's meant to try and get smart I am." It's meant to try and get you into the mindset of taking advantage you into the mindset of taking advantage you into the mindset of taking advantage of the limits that we have today. I'm of the limits that we have today. I'm of the limits that we have today. I'm currently using a Hermes agent inside of currently using a Hermes agent inside of currently using a Hermes agent inside of Discord as my main like default AI agent Discord as my main like default AI agent Discord as my main like default AI agent that does random work on my Mac mini, that does random work on my Mac mini, that does random work on my Mac mini, and I'm going to take advantage of that and I'm going to take advantage of that and I'm going to take advantage of that here. It already has Claude code set up, here. It already has Claude code set up, here. It already has Claude code set up, so watch and learn. I want to make sure so watch and learn. I want to make sure so watch and learn. I want to make sure my Claude code account has a recent my Claude code account has a recent my Claude code account has a recent message at all times. Set up a cron that message at all times. Set up a cron that message at all times. Set up a cron that triggers every 5 hours. It should run in triggers every 5 hours. It should run in triggers every 5 hours. It should run in an empty directory. The command should an empty directory. The command should an empty directory. The command should be Claude-p hi.
-
be Claude-p hi. be Claude-p hi. And now my Hermes agent is going to go And now my Hermes agent is going to go And now my Hermes agent is going to go set that up, so going forward, I will set that up, so going forward, I will set that up, so going forward, I will always have something triggered on that always have something triggered on that always have something triggered on that account to keep my rate limits moving at account to keep my rate limits moving at account to keep my rate limits moving at all times. Also, just a pro tip I got all times. Also, just a pro tip I got all times. Also, just a pro tip I got from Ben, I much prefer using open from Ben, I much prefer using open from Ben, I much prefer using open Claude and Hermes in something like Claude and Hermes in something like Claude and Hermes in something like Discord because every message gets its Discord because every message gets its Discord because every message gets its own thread that I can keep following up own thread that I can keep following up own thread that I can keep following up into maintain context. It's really, into maintain context. It's really, into maintain context. It's really, really nice. And I'm using this for a really nice. And I'm using this for a really nice. And I'm using this for a ton of different stuff. Obviously, you ton of different stuff. Obviously, you ton of different stuff. Obviously, you can't use your Claude code off directly can't use your Claude code off directly can't use your Claude code off directly inside of a tool like this because inside of a tool like this because inside of a tool like this because Anthropic really wants you to use Claude Anthropic really wants you to use Claude Anthropic really wants you to use Claude code with your sub, not tools like this, code with your sub, not tools like this, code with your sub, not tools like this, but you can use your Codex sub with this but you can use your Codex sub with this but you can use your Codex sub with this if you want to min-max that, too, which if you want to min-max that, too, which if you want to min-max that, too, which is what I do. It works really good. is what I do. It works really good. is what I do. It works really good. Fantastic. Now that's done, I'll never Fantastic. Now that's done, I'll never Fantastic. Now that's done, I'll never again have to worry about my limits not again have to worry about my limits not again have to worry about my limits not counting down. Again, this isn't trying counting down. Again, this isn't trying counting down. Again, this isn't trying to get more out of your plan. Like I'm to get more out of your plan. Like I'm to get more out of your plan. Like I'm not trying to push past my weekly limit not trying to push past my weekly limit not trying to push past my weekly limit or anything here. I'm just trying to or anything here. I'm just trying to or anything here. I'm just trying to make sure that you can get to 100% and make sure that you can get to 100% and make sure that you can get to 100% and get your reset as quickly as possible. get your reset as quickly as possible. get your reset as quickly as possible. But as I mentioned before, I'm juggling But as I mentioned before, I'm juggling But as I mentioned before, I'm juggling two accounts. This might seem like two accounts. This might seem like two accounts. This might seem like really, really annoying to do, really, really annoying to do, really, really annoying to do, especially if you're using the Claude especially if you're using the Claude especially if you're using the Claude desktop app because when you sign out desktop app because when you sign out desktop app because when you sign out and into a different account, you lose and into a different account, you lose and into a different account, you lose all of your history, all of your all of your history, all of your all of your history, all of your sessions, all of your everything. I put sessions, all of your everything. I put sessions, all of your everything. I put out a warning in that this in the Claude out a warning in that this in the Claude out a warning in that this in the Claude code desktop app for us account code desktop app for us account code desktop app for us account switching folks, switching folks, switching folks, and Anthony from Anthropic actually and Anthony from Anthropic actually and Anthony from Anthropic actually confirmed that they're going to fix this confirmed that they're going to fix this confirmed that they're going to fix this behavior soon. I don't think that will behavior soon. I don't think that will behavior soon. I don't think that will happen within the range we have Fable happen within the range we have Fable happen within the range we have Fable for in the subscriptions, so I wouldn't for in the subscriptions, so I wouldn't for in the subscriptions, so I wouldn't count on this. And honestly, the CLI count on this. And honestly, the CLI count on this. And honestly, the CLI still is a much better tool than the still is a much better tool than the still is a much better tool than the desktop app at this point in time.
-
desktop app at this point in time. desktop app at this point in time. Desktop has made meaningful improvements Desktop has made meaningful improvements Desktop has made meaningful improvements I did my video roasting it, but I I I did my video roasting it, but I I I did my video roasting it, but I I tried it a bunch last few days. I tried it a bunch last few days. I tried it a bunch last few days. I wouldn't recommend it. wouldn't recommend it. wouldn't recommend it. So, what does account switching look So, what does account switching look So, what does account switching look like when you're actually doing real like when you're actually doing real like when you're actually doing real work with these tools? I could tell you work with these tools? I could tell you work with these tools? I could tell you how to take advantage of the multiple how to take advantage of the multiple how to take advantage of the multiple accounts, but I'd rather just show you. accounts, but I'd rather just show you. accounts, but I'd rather just show you. And when I do this, I'm also going to And when I do this, I'm also going to And when I do this, I'm also going to show one of the real use cases I've been show one of the real use cases I've been show one of the real use cases I've been doing with this stuff. I've talked a doing with this stuff. I've talked a doing with this stuff. I've talked a bunch about my recent project Lakebed in bunch about my recent project Lakebed in bunch about my recent project Lakebed in various videos. I'm trying to make a various videos. I'm trying to make a various videos. I'm trying to make a better cloud for building apps that better cloud for building apps that better cloud for building apps that agents can easily operate without agents can easily operate without agents can easily operate without needing to access dashboards, deal with needing to access dashboards, deal with needing to access dashboards, deal with API keys, and all those types of things. API keys, and all those types of things. API keys, and all those types of things. And I have a lot of work in flight on And I have a lot of work in flight on And I have a lot of work in flight on this project right now. In fact, some of this project right now. In fact, some of this project right now. In fact, some of it overlaps. PRs 35, 37, and 39 are all it overlaps. PRs 35, 37, and 39 are all it overlaps. PRs 35, 37, and 39 are all attempts to add file storage to Lakebed attempts to add file storage to Lakebed attempts to add file storage to Lakebed that have their own benefits and that have their own benefits and that have their own benefits and negatives, respectively. 35 and 37 negatives, respectively. 35 and 37 negatives, respectively. 35 and 37 differed a lot in their implementations. differed a lot in their implementations. differed a lot in their implementations. I had Codex and Claude Code both I had Codex and Claude Code both I had Codex and Claude Code both building their own solutions and then building their own solutions and then building their own solutions and then comparing each others, and they couldn't comparing each others, and they couldn't comparing each others, and they couldn't have disagreed more. But when I have disagreed more. But when I have disagreed more. But when I synthesize the best parts of both, it synthesize the best parts of both, it synthesize the best parts of both, it made a new PR with 39, and it came out a made a new PR with 39, and it came out a made a new PR with 39, and it came out a lot better. I'm still not 100% lot better. I'm still not 100% lot better. I'm still not 100% confident, though, and rather than doing confident, though, and rather than doing confident, though, and rather than doing the normal thing, which is sitting there the normal thing, which is sitting there the normal thing, which is sitting there and reading all of this yourself, I'm and reading all of this yourself, I'm and reading all of this yourself, I'm going to be lazy and take advantage of going to be lazy and take advantage of going to be lazy and take advantage of my limits. I have Claude Code open work my limits. I have Claude Code open work my limits. I have Claude Code open work tree of this repo, and I'm going to tell tree of this repo, and I'm going to tell tree of this repo, and I'm going to tell it to get some work done here. Notice it to get some work done here. Notice it to get some work done here. Notice that I'm using Ultra Code. I don't that I'm using Ultra Code. I don't that I'm using Ultra Code. I don't recommend using Ultra Code and the recommend using Ultra Code and the recommend using Ultra Code and the workflow feature in Claude Code if workflow feature in Claude Code if workflow feature in Claude Code if you're trying to conserve tokens at all, you're trying to conserve tokens at all, you're trying to conserve tokens at all, but if you're trying to hit those but if you're trying to hit those but if you're trying to hit those limits, you're trying to take advantage limits, you're trying to take advantage limits, you're trying to take advantage of your remaining inference before a of your remaining inference before a of your remaining inference before a reset comes up, doing things like this reset comes up, doing things like this reset comes up, doing things like this is actually really useful and finds more is actually really useful and finds more is actually really useful and finds more information than I would have expected.
-
information than I would have expected. information than I would have expected. I'm going to Whisper flow this one cuz I I'm going to Whisper flow this one cuz I I'm going to Whisper flow this one cuz I don't feel like typing and talking at don't feel like typing and talking at don't feel like typing and talking at the same time, so you'll forgive me. I the same time, so you'll forgive me. I the same time, so you'll forgive me. I currently have three pull requests open currently have three pull requests open currently have three pull requests open on this project, 35, 37, and 39, all of on this project, 35, 37, and 39, all of on this project, 35, 37, and 39, all of which are implementing roughly the same which are implementing roughly the same which are implementing roughly the same feature, which is user-facing object feature, which is user-facing object feature, which is user-facing object storage that developers can implement in storage that developers can implement in storage that developers can implement in their Lakebed apps. I want to decide their Lakebed apps. I want to decide their Lakebed apps. I want to decide which one is the best choice to merge. which one is the best choice to merge. which one is the best choice to merge. Make a workflow where you break these Make a workflow where you break these Make a workflow where you break these PRs up and have judges review each of PRs up and have judges review each of PRs up and have judges review each of them independently to conclude and them independently to conclude and them independently to conclude and figure out which one is best and what figure out which one is best and what figure out which one is best and what pieces of each we can bring into the pieces of each we can bring into the pieces of each we can bring into the best solution. Audit all of them best solution. Audit all of them best solution. Audit all of them independently and help me come to a independently and help me come to a independently and help me come to a conclusion as to which of these PRs conclusion as to which of these PRs conclusion as to which of these PRs should be the one that we continue should be the one that we continue should be the one that we continue iterating on and merge eventually. I put iterating on and merge eventually. I put iterating on and merge eventually. I put the word workflow in here specifically the word workflow in here specifically the word workflow in here specifically to trigger it to start a workflow to trigger it to start a workflow to trigger it to start a workflow because workflows are a great way to do because workflows are a great way to do because workflows are a great way to do this type of giant bulk work. I've this type of giant bulk work. I've this type of giant bulk work. I've actually found it very nice for this actually found it very nice for this actually found it very nice for this type of thing. It's now creating the type of thing. It's now creating the type of thing. It's now creating the workflow and orchestrating all of these workflow and orchestrating all of these workflow and orchestrating all of these sub-agents to be the judges and I'll sub-agents to be the judges and I'll sub-agents to be the judges and I'll show what that looks like in Claude code show what that looks like in Claude code show what that looks like in Claude code cuz it is really cool. What I want to cuz it is really cool. What I want to cuz it is really cool. What I want to show you guys first is the account show you guys first is the account show you guys first is the account maxing that I've been doing. So here I maxing that I've been doing. So here I maxing that I've been doing. So here I have my personal Claude code account and have my personal Claude code account and have my personal Claude code account and here I have my secondary one I was here I have my secondary one I was here I have my secondary one I was mentioning before. The personal is the mentioning before. The personal is the mentioning before. The personal is the one I currently have signed in. So if I one I currently have signed in. So if I one I currently have signed in. So if I refresh, you're going to see that usage refresh, you're going to see that usage refresh, you're going to see that usage start to tick up pretty fast.
-
start to tick up pretty fast. start to tick up pretty fast. And not. And not. And not. Okay, not that fast because it's still Okay, not that fast because it's still Okay, not that fast because it's still figuring out what the workflow is. So I figuring out what the workflow is. So I figuring out what the workflow is. So I only have one thread going and honestly only have one thread going and honestly only have one thread going and honestly running Mythos 24/7 or Fable in this running Mythos 24/7 or Fable in this running Mythos 24/7 or Fable in this case 24/7 in just one thread, so you case 24/7 in just one thread, so you case 24/7 in just one thread, so you only have one going at a time, you're only have one going at a time, you're only have one going at a time, you're probably not going to hit the limits too probably not going to hit the limits too probably not going to hit the limits too aggressively. But when you start getting aggressively. But when you start getting aggressively. But when you start getting workflows going where you have eight or workflows going where you have eight or workflows going where you have eight or more running at the same time, you'll more running at the same time, you'll more running at the same time, you'll burn through the limits fast, like burn through the limits fast, like burn through the limits fast, like easily under an hour. Okay, it's still easily under an hour. Okay, it's still easily under an hour. Okay, it's still setting up the environment, but we do setting up the environment, but we do setting up the environment, but we do see the percent starting to go up. We're see the percent starting to go up. We're see the percent starting to go up. We're now at one. It'll go to two momentarily. now at one. It'll go to two momentarily. now at one. It'll go to two momentarily. Well, it would, but I want to demo the Well, it would, but I want to demo the Well, it would, but I want to demo the account swapping. account swapping. account swapping. It's a really, really complex process. It's a really, really complex process. It's a really, really complex process. I'll show you just how complex. I'll show you just how complex. I'll show you just how complex. You run Claude code. You run Claude code. You run Claude code. You run the {slash} login command. You run the {slash} login command. You run the {slash} login command. You grab the URL that it puts out and You grab the URL that it puts out and You grab the URL that it puts out and you put it in whatever browser profile you put it in whatever browser profile you put it in whatever browser profile has the Claude code that you want. I'm has the Claude code that you want. I'm has the Claude code that you want. I'm using Helium, which has browser profiles using Helium, which has browser profiles using Helium, which has browser profiles just like Chrome does, so I can swap just like Chrome does, so I can swap just like Chrome does, so I can swap between the two. The purple is for my between the two. The purple is for my between the two. The purple is for my secondary account. I authorize here. secondary account. I authorize here. secondary account. I authorize here. I go back to my terminal. I go back to my terminal. I go back to my terminal. And now I'm logged in on a different And now I'm logged in on a different And now I'm logged in on a different account. account. account. And ready for the crazy part?
-
And ready for the crazy part? And ready for the crazy part? We have this workflow going generating We have this workflow going generating We have this workflow going generating tokens. The next turn it takes, as in tokens. The next turn it takes, as in tokens. The next turn it takes, as in the next time it does a tool call or the next time it does a tool call or the next time it does a tool call or kicks off a workflow or starts a next kicks off a workflow or starts a next kicks off a workflow or starts a next step of any form, that one's going to step of any form, that one's going to step of any form, that one's going to use the new auth token that it just got use the new auth token that it just got use the new auth token that it just got and and to start routing to the other and and to start routing to the other and and to start routing to the other account. It doesn't care when you swap account. It doesn't care when you swap account. It doesn't care when you swap accounts mid-session. So, if I have five accounts mid-session. So, if I have five accounts mid-session. So, if I have five workflows going across different workflows going across different workflows going across different projects or different work trees on the projects or different work trees on the projects or different work trees on the same project, and I want to make sure same project, and I want to make sure same project, and I want to make sure that I'm not going to have them all stop that I'm not going to have them all stop that I'm not going to have them all stop immediately because I hit a limit, which immediately because I hit a limit, which immediately because I hit a limit, which is super annoying, by the way, because is super annoying, by the way, because is super annoying, by the way, because it doesn't recover these workflows well. it doesn't recover these workflows well. it doesn't recover these workflows well. So, if you have a workflow with like 100 So, if you have a workflow with like 100 So, if you have a workflow with like 100 plus sub-agents, you get to the 94th, plus sub-agents, you get to the 94th, plus sub-agents, you get to the 94th, and then your limit hits, you have to and then your limit hits, you have to and then your limit hits, you have to often rerun the entire workflow when you often rerun the entire workflow when you often rerun the entire workflow when you hit those limits. So, it's worth it to hit those limits. So, it's worth it to hit those limits. So, it's worth it to keep an eye on how close you're getting keep an eye on how close you're getting keep an eye on how close you're getting and swap your off out before you hit and swap your off out before you hit and swap your off out before you hit those limits. Or, you can do what I do, those limits. Or, you can do what I do, those limits. Or, you can do what I do, which is burn a [ __ ] load of money by which is burn a [ __ ] load of money by which is burn a [ __ ] load of money by turning on overage credits. And once you turning on overage credits. And once you turning on overage credits. And once you see those getting burned, then you see those getting burned, then you see those getting burned, then you switch over. But now you see, this switch over. But now you see, this switch over. But now you see, this account is no longer going up because account is no longer going up because account is no longer going up because this is not the account that the traffic this is not the account that the traffic this is not the account that the traffic is being routed through. Okay. Now the is being routed through. Okay. Now the is being routed through. Okay. Now the workflow is going, and you can see just workflow is going, and you can see just workflow is going, and you can see just how many tokens get burned with these how many tokens get burned with these how many tokens get burned with these things. We have the audit stage, which things. We have the audit stage, which things. We have the audit stage, which will be 13 separate agents. And will be 13 separate agents. And will be 13 separate agents. And depending on what their results are, depending on what their results are, depending on what their results are, they'll pass it on to the judge section, they'll pass it on to the judge section, they'll pass it on to the judge section, which will have even more. Then verify, which will have even more. Then verify, which will have even more. Then verify, harvest, and synthesize. This run looks harvest, and synthesize. This run looks harvest, and synthesize. This run looks like it might be 100 plus of these like it might be 100 plus of these like it might be 100 plus of these sub-agents, which is going to be crazy.
-
sub-agents, which is going to be crazy. sub-agents, which is going to be crazy. Burns absurd tokens. We're already at Burns absurd tokens. We're already at Burns absurd tokens. We're already at 368,000 tokens because it's running 368,000 tokens because it's running 368,000 tokens because it's running eight of them in parallel right now. So, eight of them in parallel right now. So, eight of them in parallel right now. So, now if we go here, you'll see these rate now if we go here, you'll see these rate now if we go here, you'll see these rate limits for my second account are going limits for my second account are going limits for my second account are going up super fast. We're already at 5% when up super fast. We're already at 5% when up super fast. We're already at 5% when it took the last 20 minutes to even hit it took the last 20 minutes to even hit it took the last 20 minutes to even hit 1% here, because now we're running it 8x 1% here, because now we're running it 8x 1% here, because now we're running it 8x harder. But now this account isn't going harder. But now this account isn't going harder. But now this account isn't going up anymore cuz it's not the one being up anymore cuz it's not the one being up anymore cuz it's not the one being used for the sub-agent run. Super used for the sub-agent run. Super used for the sub-agent run. Super helpful for hopping between accounts, helpful for hopping between accounts, helpful for hopping between accounts, and the fact that you can just do and the fact that you can just do and the fact that you can just do {slash} login in any Claude code {slash} login in any Claude code {slash} login in any Claude code terminal on a given machine, and it terminal on a given machine, and it terminal on a given machine, and it updates all of the things on that updates all of the things on that updates all of the things on that machine is wonderful. Jesus Christ, machine is wonderful. Jesus Christ, machine is wonderful. Jesus Christ, we're almost at a million tokens down, we're almost at a million tokens down, we're almost at a million tokens down, and this has been like under a minute. and this has been like under a minute. and this has been like under a minute. Yeah, this percentage is going up fast. Yeah, this percentage is going up fast. Yeah, this percentage is going up fast. Like this is real time. We're already at Like this is real time. We're already at Like this is real time. We're already at seven. seven. seven. This is why you should be careful with This is why you should be careful with This is why you should be careful with your workflows, but you should also be your workflows, but you should also be your workflows, but you should also be careful with your weekly limits. My two careful with your weekly limits. My two careful with your weekly limits. My two accounts reset on Wednesday and accounts reset on Wednesday and accounts reset on Wednesday and Thursday, respectively, right now. And Thursday, respectively, right now. And Thursday, respectively, right now. And I'm already at pretty high percentages I'm already at pretty high percentages I'm already at pretty high percentages on both. I'm expecting to max out my on both. I'm expecting to max out my on both. I'm expecting to max out my weekly on at least one of these accounts weekly on at least one of these accounts weekly on at least one of these accounts and I'm kind of counting on a reset for and I'm kind of counting on a reset for and I'm kind of counting on a reset for some reason. Hopeful that they'll have some reason. Hopeful that they'll have some reason. Hopeful that they'll have some reason to reset things, but even if some reason to reset things, but even if some reason to reset things, but even if they don't, I'll probably just grab they don't, I'll probably just grab they don't, I'll probably just grab another $200 account and push it to its another $200 account and push it to its another $200 account and push it to its limits as well, specifically during this limits as well, specifically during this limits as well, specifically during this short testing window of 10 days where we short testing window of 10 days where we short testing window of 10 days where we can use Fable. I will not be going can use Fable. I will not be going can use Fable. I will not be going anywhere near this hard on Anthropic anywhere near this hard on Anthropic anywhere near this hard on Anthropic models when Fable is no longer in the models when Fable is no longer in the models when Fable is no longer in the subscription tiers, but when it is, I'm subscription tiers, but when it is, I'm subscription tiers, but when it is, I'm pushing my limits. So, you might be pushing my limits. So, you might be pushing my limits. So, you might be wondering now, how big are those weekly wondering now, how big are those weekly wondering now, how big are those weekly limits? Well, from my experience, just limits? Well, from my experience, just limits? Well, from my experience, just from basic testing in this secondary from basic testing in this secondary from basic testing in this secondary account in particular, it seems like you account in particular, it seems like you account in particular, it seems like you can get roughly 25% of your weekly when can get roughly 25% of your weekly when can get roughly 25% of your weekly when you hit 100% of your 5-hour. Put simply,
-
you hit 100% of your 5-hour. Put simply, you hit 100% of your 5-hour. Put simply, you can max out a 5-hour window four you can max out a 5-hour window four you can max out a 5-hour window four times in a given week before you're out times in a given week before you're out times in a given week before you're out of usage. So, should you always be of usage. So, should you always be of usage. So, should you always be trying to hit the 5-hour limits? Maybe, trying to hit the 5-hour limits? Maybe, trying to hit the 5-hour limits? Maybe, especially during the end, but you especially during the end, but you especially during the end, but you should absolutely be aiming to hit the should absolutely be aiming to hit the should absolutely be aiming to hit the weekly limit to get the most out of your weekly limit to get the most out of your weekly limit to get the most out of your usage. So, I've already burned like $400 usage. So, I've already burned like $400 usage. So, I've already burned like $400 of inference just with these basic tests of inference just with these basic tests of inference just with these basic tests as I've been filming, and you might be as I've been filming, and you might be as I've been filming, and you might be thinking, "Wow, that's a lot of thinking, "Wow, that's a lot of thinking, "Wow, that's a lot of inference for things that haven't inference for things that haven't inference for things that haven't actually panned out much yet. Like actually panned out much yet. Like actually panned out much yet. Like you're reviewing three slot PRs and you're reviewing three slot PRs and you're reviewing three slot PRs and trying to decide which one is best. trying to decide which one is best. trying to decide which one is best. Like, how valuable is that?" First off, Like, how valuable is that?" First off, Like, how valuable is that?" First off, that is actually quite valuable when that is actually quite valuable when that is actually quite valuable when you're trying to figure out how to deal you're trying to figure out how to deal you're trying to figure out how to deal with a giant pile of PRs, but second with a giant pile of PRs, but second with a giant pile of PRs, but second off, you can modify this slightly to be off, you can modify this slightly to be off, you can modify this slightly to be way more useful. Here's a much more way more useful. Here's a much more way more useful. Here's a much more realistic example that I find is realistic example that I find is realistic example that I find is actually super useful in my day-to-day actually super useful in my day-to-day actually super useful in my day-to-day work. I maintain a handful of different work. I maintain a handful of different work. I maintain a handful of different repos. From Lakebed, which still isn't repos. From Lakebed, which still isn't repos. From Lakebed, which still isn't public yet, to T3 Chat, which also isn't public yet, to T3 Chat, which also isn't public yet, to T3 Chat, which also isn't a public repo, but at the very least has a public repo, but at the very least has a public repo, but at the very least has real people contributing every day, to real people contributing every day, to real people contributing every day, to T3 Code, which is a big open source T3 Code, which is a big open source T3 Code, which is a big open source project that has a lot of people project that has a lot of people project that has a lot of people throwing stuff at it. Even with our best throwing stuff at it. Even with our best throwing stuff at it. Even with our best efforts, keeping it under 400 issues and efforts, keeping it under 400 issues and efforts, keeping it under 400 issues and under 300 PRs feels nearly impossible.
-
under 300 PRs feels nearly impossible. under 300 PRs feels nearly impossible. So, rather than try to do that, I've So, rather than try to do that, I've So, rather than try to do that, I've been putting more effort into been putting more effort into been putting more effort into highlighting and surfacing the best work highlighting and surfacing the best work highlighting and surfacing the best work that is worth pursuing. Reading all that is worth pursuing. Reading all that is worth pursuing. Reading all those PRs and keeping up with all those those PRs and keeping up with all those those PRs and keeping up with all those changes is way too much work for any changes is way too much work for any changes is way too much work for any reasonable human to even consider reasonable human to even consider reasonable human to even consider trying, which is why I don't. I let my trying, which is why I don't. I let my trying, which is why I don't. I let my agents do it because they don't get agents do it because they don't get agents do it because they don't get bored. They can just sit there and do bored. They can just sit there and do bored. They can just sit there and do the thing indefinitely. So, one of the the thing indefinitely. So, one of the the thing indefinitely. So, one of the things I have Mythos doing for me, well, things I have Mythos doing for me, well, things I have Mythos doing for me, well, Fable, same difference, is every morning Fable, same difference, is every morning Fable, same difference, is every morning going through all of my PRs on all of my going through all of my PRs on all of my going through all of my PRs on all of my repos and helping me surface the ones repos and helping me surface the ones repos and helping me surface the ones that are the easiest to merge, the most that are the easiest to merge, the most that are the easiest to merge, the most justifiable to get done, etc. This justifiable to get done, etc. This justifiable to get done, etc. This example is using my Hermes agent on top example is using my Hermes agent on top example is using my Hermes agent on top of my Codex sub, but this is the one of my Codex sub, but this is the one of my Codex sub, but this is the one that Mythos made and it's really good. that Mythos made and it's really good. that Mythos made and it's really good. I ran this one on just T3 Code and it I ran this one on just T3 Code and it I ran this one on just T3 Code and it built this ranked Q section where it built this ranked Q section where it built this ranked Q section where it goes through every PR that's currently goes through every PR that's currently goes through every PR that's currently open, gives it a status, and ranks them open, gives it a status, and ranks them open, gives it a status, and ranks them based on how easy is this to just go based on how easy is this to just go based on how easy is this to just go merge, and how much of my attention does merge, and how much of my attention does merge, and how much of my attention does it deserve. So, number one here we have it deserve. So, number one here we have it deserve. So, number one here we have disabling external Git diffs, which is a disabling external Git diffs, which is a disabling external Git diffs, which is a PR that Magnus opened to fix a bug with PR that Magnus opened to fix a bug with PR that Magnus opened to fix a bug with the external diff viewer some people use the external diff viewer some people use the external diff viewer some people use with Git when they're using T3 Code. with Git when they're using T3 Code. with Git when they're using T3 Code. Simple PR, fixes a real bug, people will Simple PR, fixes a real bug, people will Simple PR, fixes a real bug, people will be very happy if we merge this. It was be very happy if we merge this. It was be very happy if we merge this. It was surfaced by the agent even though this surfaced by the agent even though this surfaced by the agent even though this PR was originally filed PR was originally filed PR was originally filed And as soon as I sent this link with all And as soon as I sent this link with all And as soon as I sent this link with all of these ranked PRs to Julius, it got of these ranked PRs to Julius, it got of these ranked PRs to Julius, it got merged within 5 minutes. Getting this merged within 5 minutes. Getting this merged within 5 minutes. Getting this type of overview of all of your work is type of overview of all of your work is type of overview of all of your work is so much better than trying to dig so much better than trying to dig so much better than trying to dig through the GitHub PR tab trying to get through the GitHub PR tab trying to get through the GitHub PR tab trying to get useful information out of it. If you're useful information out of it. If you're useful information out of it. If you're wondering how I got my agent to spit out wondering how I got my agent to spit out wondering how I got my agent to spit out these HTML plans, that's another service these HTML plans, that's another service these HTML plans, that's another service that I threw together. This one's just that I threw together. This one's just that I threw together. This one's just for me and my team, but I recommend for me and my team, but I recommend for me and my team, but I recommend building something like it and maybe building something like it and maybe building something like it and maybe I'll even open source it in the future I'll even open source it in the future I'll even open source it in the future if enough people want it. I built the
-
if enough people want it. I built the if enough people want it. I built the simple service for hosting plan files, simple service for hosting plan files, simple service for hosting plan files, which are just HTML plan descriptions, which are just HTML plan descriptions, which are just HTML plan descriptions, on a real web service. So, my agent can on a real web service. So, my agent can on a real web service. So, my agent can spit out a URL that I can click on and spit out a URL that I can click on and spit out a URL that I can click on and then go see what it was thinking or what then go see what it was thinking or what then go see what it was thinking or what it's planning. I find HTML plans to be it's planning. I find HTML plans to be it's planning. I find HTML plans to be way more readable than markdown. I have way more readable than markdown. I have way more readable than markdown. I have a whole video about this, and I've been a whole video about this, and I've been a whole video about this, and I've been abusing these for all sorts of different abusing these for all sorts of different abusing these for all sorts of different things, including reviewing agent work things, including reviewing agent work things, including reviewing agent work and having one agent review another and having one agent review another and having one agent review another agent's work and then just pass the HTML agent's work and then just pass the HTML agent's work and then just pass the HTML over, but also having the ability to over, but also having the ability to over, but also having the ability to read it in a good format. That's what's read it in a good format. That's what's read it in a good format. That's what's cool about HTML is I can look at it in cool about HTML is I can look at it in cool about HTML is I can look at it in my browser or on my phone in my browser my browser or on my phone in my browser my browser or on my phone in my browser and get a good idea of what's going on, and get a good idea of what's going on, and get a good idea of what's going on, and then just paste the link to an agent and then just paste the link to an agent and then just paste the link to an agent and say, "Hey, go deal with this." And and say, "Hey, go deal with this." And and say, "Hey, go deal with this." And it figures it out. By the way, that it figures it out. By the way, that it figures it out. By the way, that agent run I started earlier with the agent run I started earlier with the agent run I started earlier with the workflow is now at 1.6 million tokens workflow is now at 1.6 million tokens workflow is now at 1.6 million tokens and 21% of this fresh usage window used. and 21% of this fresh usage window used. and 21% of this fresh usage window used. And that is in under 30 minutes. And that is in under 30 minutes. And that is in under 30 minutes. Kind of crazy. You can burn through Kind of crazy. You can burn through Kind of crazy. You can burn through usage fast, but if you steer it in the usage fast, but if you steer it in the usage fast, but if you steer it in the right way, you can get actually useful right way, you can get actually useful right way, you can get actually useful stuff. I actually did have PR 39 get stuff. I actually did have PR 39 get stuff. I actually did have PR 39 get reviewed previously by Mythos. And I reviewed previously by Mythos. And I reviewed previously by Mythos. And I have its review here, which is super have its review here, which is super have its review here, which is super convenient because Codex was actually convenient because Codex was actually convenient because Codex was actually the agent that wrote this PR. So, having the agent that wrote this PR. So, having the agent that wrote this PR. So, having Codex write it, having Mythos review it, Codex write it, having Mythos review it, Codex write it, having Mythos review it, and I can take this URL, hand it to and I can take this URL, hand it to and I can take this URL, hand it to Codex and say, "Hey, go make the changes Codex and say, "Hey, go make the changes Codex and say, "Hey, go make the changes that this suggested and give me feedback that this suggested and give me feedback that this suggested and give me feedback on the ones you don't like as much."
-
on the ones you don't like as much." on the ones you don't like as much." It's such a useful way to just pass It's such a useful way to just pass It's such a useful way to just pass context around in a way I can read. So, context around in a way I can read. So, context around in a way I can read. So, yeah. Again, points in favor of HTML yeah. Again, points in favor of HTML yeah. Again, points in favor of HTML plans, highly recommend it. If you're plans, highly recommend it. If you're plans, highly recommend it. If you're already pushing the limits of workflows already pushing the limits of workflows already pushing the limits of workflows and reviews and you're still not getting and reviews and you're still not getting and reviews and you're still not getting close to those limits, there are plenty close to those limits, there are plenty close to those limits, there are plenty more things you can do, don't worry. One more things you can do, don't worry. One more things you can do, don't worry. One that I would highly recommend is looking that I would highly recommend is looking that I would highly recommend is looking into the skills others have made to into the skills others have made to into the skills others have made to audit your projects and your code and audit your projects and your code and audit your projects and your code and find ways to improve it. For example, find ways to improve it. For example, find ways to improve it. For example, Shaden's Shaden improve skill is really Shaden's Shaden improve skill is really Shaden's Shaden improve skill is really nice. Vercel built this nice plugin for nice. Vercel built this nice plugin for nice. Vercel built this nice plugin for adding skills. If you want to actually adding skills. If you want to actually adding skills. If you want to actually add the skills, there's cool commands add the skills, there's cool commands add the skills, there's cool commands like the Vercel skills package, but you like the Vercel skills package, but you like the Vercel skills package, but you can also just go copy the content of the can also just go copy the content of the can also just go copy the content of the skill directly. Because as you hopefully skill directly. Because as you hopefully skill directly. Because as you hopefully know by now, skills are for the most know by now, skills are for the most know by now, skills are for the most part just markdown. part just markdown. part just markdown. So, here we have all the markdown for So, here we have all the markdown for So, here we have all the markdown for this skill. I can go to raw. I can grab this skill. I can go to raw. I can grab this skill. I can go to raw. I can grab this all. I can open up Claude. I could this all. I can open up Claude. I could this all. I can open up Claude. I could put it in workflow mode or ultra code put it in workflow mode or ultra code put it in workflow mode or ultra code mode, but I am already burning enough mode, but I am already burning enough mode, but I am already burning enough inference and I want to save some of inference and I want to save some of inference and I want to save some of this for later. So, I'm just going to this for later. So, I'm just going to this for later. So, I'm just going to paste, enter, and let it do its thing. paste, enter, and let it do its thing. paste, enter, and let it do its thing. Again, very useful if you notice Again, very useful if you notice Again, very useful if you notice yourself at the end of a weekly limit or yourself at the end of a weekly limit or yourself at the end of a weekly limit or an hourly limit and you want to get a an hourly limit and you want to get a an hourly limit and you want to get a little bit more inference out, just keep little bit more inference out, just keep little bit more inference out, just keep a set of these things in your head that a set of these things in your head that a set of these things in your head that you would like to run and write them you would like to run and write them you would like to run and write them down, maybe even queue them. I've also down, maybe even queue them. I've also down, maybe even queue them. I've also I've been loving Raycast's notes I've been loving Raycast's notes I've been loving Raycast's notes feature. It's super easy to just like feature. It's super easy to just like feature. It's super easy to just like write something down write something down write something down and then hide and reopen it with a hot and then hide and reopen it with a hot and then hide and reopen it with a hot key. And you can easily iterate through key. And you can easily iterate through key. And you can easily iterate through the notes that you have saved here to the notes that you have saved here to the notes that you have saved here to find random prompts or things you want find random prompts or things you want find random prompts or things you want to do and go grab them in the future if to do and go grab them in the future if to do and go grab them in the future if you don't want to send them just yet.
-
you don't want to send them just yet. you don't want to send them just yet. Looks like it's going to kick off a Looks like it's going to kick off a Looks like it's going to kick off a workflow for the improving anyways, workflow for the improving anyways, workflow for the improving anyways, which means I'm about to burn a lot of which means I'm about to burn a lot of which means I'm about to burn a lot of money. money. money. Great. Thankfully, it's not my money. Great. Thankfully, it's not my money. Great. Thankfully, it's not my money. And since starting this video recording, And since starting this video recording, And since starting this video recording, we've already done another like $400 of we've already done another like $400 of we've already done another like $400 of inference. Jesus Christ. Take advantage inference. Jesus Christ. Take advantage inference. Jesus Christ. Take advantage of these generous limits while we have of these generous limits while we have of these generous limits while we have them, folks. We got to escape that them, folks. We got to escape that them, folks. We got to escape that permanent underclass. permanent underclass. permanent underclass. One thing I touched on earlier that I One thing I touched on earlier that I One thing I touched on earlier that I haven't really dug into yet is the haven't really dug into yet is the haven't really dug into yet is the effect this has had on my sleep. I'm effect this has had on my sleep. I'm effect this has had on my sleep. I'm sure we've all been there, glued to our sure we've all been there, glued to our sure we've all been there, glued to our laptop just waiting to get the results laptop just waiting to get the results laptop just waiting to get the results so you can run the next prompt and then so you can run the next prompt and then so you can run the next prompt and then go to bed. And then the one more prompt go to bed. And then the one more prompt go to bed. And then the one more prompt effect keeps you going and not leaving effect keeps you going and not leaving effect keeps you going and not leaving your desk again and again until your desk again and again until your desk again and again until eventually you're like falling asleep at eventually you're like falling asleep at eventually you're like falling asleep at your keyboard. The AI vampires of the your keyboard. The AI vampires of the your keyboard. The AI vampires of the Silicon Valley is a trope that I'm Silicon Valley is a trope that I'm Silicon Valley is a trope that I'm seeing more and more. Even some of the seeing more and more. Even some of the seeing more and more. Even some of the most skilled maintainers that are most skilled maintainers that are most skilled maintainers that are fathers of loving families are finding fathers of loving families are finding fathers of loving families are finding themselves staying up till 4:00 in the themselves staying up till 4:00 in the themselves staying up till 4:00 in the morning prompting because it's so morning prompting because it's so morning prompting because it's so addicting. I don't have a solution to addicting. I don't have a solution to addicting. I don't have a solution to that. I'm not going to sit here and that. I'm not going to sit here and that. I'm not going to sit here and pretend I do. What I do have is a Mac pretend I do. What I do have is a Mac pretend I do. What I do have is a Mac mini on the same network that has Codex, mini on the same network that has Codex, mini on the same network that has Codex, Claude code, and Hermes agent all set up Claude code, and Hermes agent all set up Claude code, and Hermes agent all set up and ready to go on it, which has been and ready to go on it, which has been and ready to go on it, which has been wonderful for being able to get [ __ ] wonderful for being able to get [ __ ] wonderful for being able to get [ __ ] done from my phone and from other done from my phone and from other done from my phone and from other computers and most importantly, being computers and most importantly, being computers and most importantly, being able to close the lid on my laptop and able to close the lid on my laptop and able to close the lid on my laptop and not worry about the work stopping. I not worry about the work stopping. I not worry about the work stopping. I have three ways I interface with that have three ways I interface with that have three ways I interface with that Mac mini. The one that I admittedly use Mac mini. The one that I admittedly use Mac mini. The one that I admittedly use the most is the SSH directly into it.
-
the most is the SSH directly into it. the most is the SSH directly into it. When I'm on the same network, it's very When I'm on the same network, it's very When I'm on the same network, it's very easy to SSH into. When I'm not, I rely easy to SSH into. When I'm not, I rely easy to SSH into. When I'm not, I rely on Tailscale to do it, which has been on Tailscale to do it, which has been on Tailscale to do it, which has been much better than I was expecting. As an much better than I was expecting. As an much better than I was expecting. As an old WireGuard fan, seeing Tailscale in old WireGuard fan, seeing Tailscale in old WireGuard fan, seeing Tailscale in its current state is awesome. Highly its current state is awesome. Highly its current state is awesome. Highly recommend. I highly recommend using a recommend. I highly recommend using a recommend. I highly recommend using a terminal like tmux, which has both the terminal like tmux, which has both the terminal like tmux, which has both the sidebar and tabs, so you can pin your sidebar and tabs, so you can pin your sidebar and tabs, so you can pin your Mac mini or whatever other computer you Mac mini or whatever other computer you Mac mini or whatever other computer you have over SSH as its own section and have over SSH as its own section and have over SSH as its own section and easily manage and navigate it. It's been easily manage and navigate it. It's been easily manage and navigate it. It's been super nice. Since I'm on a Mac and the super nice. Since I'm on a Mac and the super nice. Since I'm on a Mac and the remote computer is a Mac, I can use the remote computer is a Mac, I can use the remote computer is a Mac, I can use the built-in screen share utility, which built-in screen share utility, which built-in screen share utility, which honestly I didn't even know about until honestly I didn't even know about until honestly I didn't even know about until recently, which makes it super easy to recently, which makes it super easy to recently, which makes it super easy to access the computer, have native hot access the computer, have native hot access the computer, have native hot keys, and all the other things you would keys, and all the other things you would keys, and all the other things you would expect when you're using this computer expect when you're using this computer expect when you're using this computer remotely, which has been again a remotely, which has been again a remotely, which has been again a lifesaver for doing this type of thing. lifesaver for doing this type of thing. lifesaver for doing this type of thing. This also works over Tailscale, which This also works over Tailscale, which This also works over Tailscale, which let me configure some things that were let me configure some things that were let me configure some things that were blocking when I was away from my blocking when I was away from my blocking when I was away from my computer. But, none of those are my computer. But, none of those are my computer. But, none of those are my favorite way to access that remote favorite way to access that remote favorite way to access that remote machine, and this is where I will admit machine, and this is where I will admit machine, and this is where I will admit we're about to get into a bit of a we're about to get into a bit of a we're about to get into a bit of a self-plug, but I hope you guys self-plug, but I hope you guys self-plug, but I hope you guys appreciate it because I think this is appreciate it because I think this is appreciate it because I think this is the coolest thing ever. Julius has put a the coolest thing ever. Julius has put a the coolest thing ever. Julius has put a ton of work into the T3 Code Remote ton of work into the T3 Code Remote ton of work into the T3 Code Remote System, where you can control a T3 Code System, where you can control a T3 Code System, where you can control a T3 Code instance from another computer. I instance from another computer. I instance from another computer. I usually do it by just connecting usually do it by just connecting usually do it by just connecting directly to the machine that I have T3 directly to the machine that I have T3 directly to the machine that I have T3 Code set up on, but there's lots of Code set up on, but there's lots of Code set up on, but there's lots of other options, too.
-
other options, too. other options, too. If you have T3 Code installed on one If you have T3 Code installed on one If you have T3 Code installed on one machine, it's really easy to give access machine, it's really easy to give access machine, it's really easy to give access to other ones. to other ones. to other ones. You can create a link or expose it over You can create a link or expose it over You can create a link or expose it over something like Tailscale, and then something like Tailscale, and then something like Tailscale, and then access it through that link or by going access it through that link or by going access it through that link or by going to the app.t3.gg code site and creating to the app.t3.gg code site and creating to the app.t3.gg code site and creating the connection there. Julius is also the connection there. Julius is also the connection there. Julius is also deep in the building of T3 Connect, deep in the building of T3 Connect, deep in the building of T3 Connect, which is a new method to connect to a which is a new method to connect to a which is a new method to connect to a remote T3 Code instance that you have on remote T3 Code instance that you have on remote T3 Code instance that you have on one of your machines from another one of your machines from another one of your machines from another machine or even your phone with the machine or even your phone with the machine or even your phone with the upcoming T3 Code app. Very excited about upcoming T3 Code app. Very excited about upcoming T3 Code app. Very excited about everything we're cooking there, but I'll everything we're cooking there, but I'll everything we're cooking there, but I'll be real, you can build all this be real, you can build all this be real, you can build all this yourself. CodeX has some of it built in yourself. CodeX has some of it built in yourself. CodeX has some of it built in already. Cloud Code pretends to, but it already. Cloud Code pretends to, but it already. Cloud Code pretends to, but it never works when I try. There's lots of never works when I try. There's lots of never works when I try. There's lots of ways to control your agents remotely, ways to control your agents remotely, ways to control your agents remotely, and I highly recommend finding methods and I highly recommend finding methods and I highly recommend finding methods like this because you'll start pushing like this because you'll start pushing like this because you'll start pushing the length of your jobs a ton when it's the length of your jobs a ton when it's the length of your jobs a ton when it's no longer locking you to your laptop. no longer locking you to your laptop. no longer locking you to your laptop. Before I had this remote Mac mini set Before I had this remote Mac mini set Before I had this remote Mac mini set up, I found myself running shorter jobs up, I found myself running shorter jobs up, I found myself running shorter jobs so I wouldn't have to worry about so I wouldn't have to worry about so I wouldn't have to worry about closing my laptop because I didn't want closing my laptop because I didn't want closing my laptop because I didn't want to be one of those people walking around to be one of those people walking around to be one of those people walking around with a half-open laptop. I really with a half-open laptop. I really with a half-open laptop. I really didn't. Once I had the Mac mini set up didn't. Once I had the Mac mini set up didn't. Once I had the Mac mini set up and I could run things from there, I and I could run things from there, I and I could run things from there, I found myself using agents entirely found myself using agents entirely found myself using agents entirely differently, letting them go off on long differently, letting them go off on long differently, letting them go off on long exploratory journeys, not expecting to exploratory journeys, not expecting to exploratory journeys, not expecting to be able to use the results, but be able to use the results, but be able to use the results, but expecting to have interesting enough expecting to have interesting enough expecting to have interesting enough findings to talk about to my agents or findings to talk about to my agents or findings to talk about to my agents or even to my other co-workers when we even to my other co-workers when we even to my other co-workers when we build these types of things. I don't build these types of things. I don't build these types of things. I don't care what method you end up using to care what method you end up using to care what method you end up using to control these remote machines, you could control these remote machines, you could control these remote machines, you could even build your own, which is a fun way even build your own, which is a fun way even build your own, which is a fun way to both burn tokens and really refine to both burn tokens and really refine to both burn tokens and really refine your workflow to fit your specific needs your workflow to fit your specific needs your workflow to fit your specific needs and expectations. If you take anything and expectations. If you take anything and expectations. If you take anything from this video, I really want you to from this video, I really want you to from this video, I really want you to take home the creativity you can apply take home the creativity you can apply take home the creativity you can apply to how you burn your tokens now. You can
-
to how you burn your tokens now. You can to how you burn your tokens now. You can have agents reviewing each other's work have agents reviewing each other's work have agents reviewing each other's work and you can even automate it. Something and you can even automate it. Something and you can even automate it. Something like Claude code is capable of doing like Claude code is capable of doing like Claude code is capable of doing babysitting with built-in loops. So, you babysitting with built-in loops. So, you babysitting with built-in loops. So, you can tell one agent to make a PR and can tell one agent to make a PR and can tell one agent to make a PR and watch whenever a new comment comes in watch whenever a new comment comes in watch whenever a new comment comes in and then tell another agent to watch a and then tell another agent to watch a and then tell another agent to watch a PR and whenever a new push happens, give PR and whenever a new push happens, give PR and whenever a new push happens, give it a bunch of feedback and leave a it a bunch of feedback and leave a it a bunch of feedback and leave a review. Now, you have two agents that review. Now, you have two agents that review. Now, you have two agents that aren't even aware of each other giving aren't even aware of each other giving aren't even aware of each other giving each other what they need to keep each other what they need to keep each other what they need to keep pushing work forward before humans even pushing work forward before humans even pushing work forward before humans even involved. You can give agents stuff like involved. You can give agents stuff like involved. You can give agents stuff like browser use. This is much stronger on browser use. This is much stronger on browser use. This is much stronger on the Codex side, but Claude code is the Codex side, but Claude code is the Codex side, but Claude code is starting to get there as well. This starting to get there as well. This starting to get there as well. This allows an agent to spin up the changes allows an agent to spin up the changes allows an agent to spin up the changes it made, record the screen showing that it made, record the screen showing that it made, record the screen showing that the changes work and then send you that the changes work and then send you that the changes work and then send you that it in the PR when they're done, not it in the PR when they're done, not it in the PR when they're done, not bothering you until they have video bothering you until they have video bothering you until they have video evidence that the work actually worked. evidence that the work actually worked. evidence that the work actually worked. As always, Pete is far ahead of the game As always, Pete is far ahead of the game As always, Pete is far ahead of the game and he has awesome examples. Codex has and he has awesome examples. Codex has and he has awesome examples. Codex has the ability to spin up threads by the ability to spin up threads by the ability to spin up threads by itself, so it's not just sub agents. It itself, so it's not just sub agents. It itself, so it's not just sub agents. It can trigger new threads and make new can trigger new threads and make new can trigger new threads and make new work trees and make real changes through work trees and make real changes through work trees and make real changes through that. So, he set up a really simple loop that. So, he set up a really simple loop that. So, he set up a really simple loop here. Tell Codex to maintain your repos. here. Tell Codex to maintain your repos. here. Tell Codex to maintain your repos. Wake up every 5 minutes and direct work Wake up every 5 minutes and direct work Wake up every 5 minutes and direct work to threads. Makes it easy to parallelize to threads. Makes it easy to parallelize to threads. Makes it easy to parallelize and steer work as needed. Since each and steer work as needed. Since each and steer work as needed. Since each thing that it thinks should be going on thing that it thinks should be going on thing that it thinks should be going on gets its own thread, it's very easy to gets its own thread, it's very easy to gets its own thread, it's very easy to hop into a given thread and tell it, hop into a given thread and tell it, hop into a given thread and tell it, "Hey, that's wrong. Go do it the other "Hey, that's wrong. Go do it the other "Hey, that's wrong. Go do it the other way." or just stop it if it doesn't make way." or just stop it if it doesn't make way." or just stop it if it doesn't make sense. If you combine this with sense. If you combine this with sense. If you combine this with something like your Sentry bug tracking something like your Sentry bug tracking something like your Sentry bug tracking or your DataDog analytics, you can or your DataDog analytics, you can or your DataDog analytics, you can combine it and give all of the resources combine it and give all of the resources combine it and give all of the resources and data available to you as the and data available to you as the and data available to you as the engineer to your agents, which let them engineer to your agents, which let them engineer to your agents, which let them find information that's useful, find information that's useful, find information that's useful, self-improve and just push things self-improve and just push things self-improve and just push things forward in a way that's really powerful
-
forward in a way that's really powerful forward in a way that's really powerful and surprisingly valuable. Jesus Christ, and surprisingly valuable. Jesus Christ, and surprisingly valuable. Jesus Christ, that workflow is now at 1.8 million that workflow is now at 1.8 million that workflow is now at 1.8 million tokens. We haven't even started the tokens. We haven't even started the tokens. We haven't even started the judges yet cuz we got one more of these judges yet cuz we got one more of these judges yet cuz we got one more of these audit agents going. audit agents going. audit agents going. I know it's like I'm lighting money on I know it's like I'm lighting money on I know it's like I'm lighting money on fire here, but I've already paid the fire here, but I've already paid the fire here, but I've already paid the money and it's actually really fun to money and it's actually really fun to money and it's actually really fun to watch. I didn't think I'd ever be this watch. I didn't think I'd ever be this watch. I didn't think I'd ever be this type. I've always been the like type. I've always been the like type. I've always been the like single-threaded, just get the thing done single-threaded, just get the thing done single-threaded, just get the thing done guy. guy. guy. I'm having a lot of fun right now. I've I'm having a lot of fun right now. I've I'm having a lot of fun right now. I've been going a little further than I been going a little further than I been going a little further than I normally do and I really like how Sawyer normally do and I really like how Sawyer normally do and I really like how Sawyer framed this. You need to be more framed this. You need to be more framed this. You need to be more ambitious than you have been before. Ask ambitious than you have been before. Ask ambitious than you have been before. Ask the model to rewrite your entire the model to rewrite your entire the model to rewrite your entire production app from scratch. Ask it to production app from scratch. Ask it to production app from scratch. Ask it to make an entire product or internal tool make an entire product or internal tool make an entire product or internal tool for you, but don't stop there. Ask it to for you, but don't stop there. Ask it to for you, but don't stop there. Ask it to deploy it and add accounts, multiplayer, deploy it and add accounts, multiplayer, deploy it and add accounts, multiplayer, etc. All the things you wouldn't etc. All the things you wouldn't etc. All the things you wouldn't normally have in a throwaway personal normally have in a throwaway personal normally have in a throwaway personal app. Raise your bar. Only by pushing app. Raise your bar. Only by pushing app. Raise your bar. Only by pushing more do you learn. Orchestrate. When more do you learn. Orchestrate. When more do you learn. Orchestrate. When coding and having an agent write a plan, coding and having an agent write a plan, coding and having an agent write a plan, have another agent or multiple validate have another agent or multiple validate have another agent or multiple validate each claim in that plan. Quad workflows each claim in that plan. Quad workflows each claim in that plan. Quad workflows are [ __ ] amazing for this. Sub-agents are [ __ ] amazing for this. Sub-agents are [ __ ] amazing for this. Sub-agents in Codex are getting there now, too. in Codex are getting there now, too. in Codex are getting there now, too. They're not quite as good, but they're They're not quite as good, but they're They're not quite as good, but they're better than I thought they would be. better than I thought they would be. better than I thought they would be. Fable's really good at orchestration. Fable's really good at orchestration. Fable's really good at orchestration. One tip with that though, if you don't One tip with that though, if you don't One tip with that though, if you don't want to burn too much, explicitly tell want to burn too much, explicitly tell want to burn too much, explicitly tell Fable during orchestration to use Opus Fable during orchestration to use Opus Fable during orchestration to use Opus or Sonnet for its sub-agents because for or Sonnet for its sub-agents because for or Sonnet for its sub-agents because for whatever reason, Fable, more than any whatever reason, Fable, more than any whatever reason, Fable, more than any other model, really likes using Fable other model, really likes using Fable other model, really likes using Fable for its sub-agents. It seems to not for its sub-agents. It seems to not for its sub-agents. It seems to not trust other models. So, be explicit.
-
trust other models. So, be explicit. trust other models. So, be explicit. Tell it to do other things. Also, see Tell it to do other things. Also, see Tell it to do other things. Also, see there? It just finished the audit stage there? It just finished the audit stage there? It just finished the audit stage and then it reviewed what it got back and then it reviewed what it got back and then it reviewed what it got back and decided it needed seven judge agents and decided it needed seven judge agents and decided it needed seven judge agents to do more work there, as well as these to do more work there, as well as these to do more work there, as well as these harvest agents that spun up in the harvest agents that spun up in the harvest agents that spun up in the background, as well. But again, Fable's background, as well. But again, Fable's background, as well. But again, Fable's incredible at orchestration. Fable is incredible at orchestration. Fable is incredible at orchestration. Fable is not needed for the individual not needed for the individual not needed for the individual sub-agents. If you're not trying to sub-agents. If you're not trying to sub-agents. If you're not trying to actually burn as much money as possible, actually burn as much money as possible, actually burn as much money as possible, maybe tell Fable to use something else. maybe tell Fable to use something else. maybe tell Fable to use something else. Back to what Sawyer said here. Have it Back to what Sawyer said here. Have it Back to what Sawyer said here. Have it use Codex-P to delegate parts of it to use Codex-P to delegate parts of it to use Codex-P to delegate parts of it to Codex. Have it use Opus sub-agents. Have Codex. Have it use Opus sub-agents. Have Codex. Have it use Opus sub-agents. Have it use workflows. Don't be afraid to it use workflows. Don't be afraid to it use workflows. Don't be afraid to give those models big chunks of work. give those models big chunks of work. give those models big chunks of work. Give your agent a browser and computer Give your agent a browser and computer Give your agent a browser and computer use. Let them click around and take use. Let them click around and take use. Let them click around and take screenshots. Use the JS debugger and screenshots. Use the JS debugger and screenshots. Use the JS debugger and take performance profiles. LLMs aren't take performance profiles. LLMs aren't take performance profiles. LLMs aren't just coding machines. They can do the just coding machines. They can do the just coding machines. They can do the debugging and reproducing work as well. debugging and reproducing work as well. debugging and reproducing work as well. For token maxing, get some way to kick For token maxing, get some way to kick For token maxing, get some way to kick off coding tasks from your phone. Could off coding tasks from your phone. Could off coding tasks from your phone. Could be the Codex app, ChatGPT, Claude, etc. be the Codex app, ChatGPT, Claude, etc. be the Codex app, ChatGPT, Claude, etc. As I mentioned before, I threw one of As I mentioned before, I threw one of As I mentioned before, I threw one of these in my Discord with my Hermes these in my Discord with my Hermes these in my Discord with my Hermes agent. Whatever you choose to do doesn't agent. Whatever you choose to do doesn't agent. Whatever you choose to do doesn't matter. Make it easy to get the idea out matter. Make it easy to get the idea out matter. Make it easy to get the idea out of your head into an agent running. I of your head into an agent running. I of your head into an agent running. I don't care how you set this up. There's don't care how you set this up. There's don't care how you set this up. There's lots of good options, but your goal lots of good options, but your goal lots of good options, but your goal should be to make it as easy as possible should be to make it as easy as possible should be to make it as easy as possible to go from, "Oh, that's a nice idea." to to go from, "Oh, that's a nice idea." to to go from, "Oh, that's a nice idea." to having something running, testing out having something running, testing out having something running, testing out that theory for you as quickly and that theory for you as quickly and that theory for you as quickly and smoothly as possible. The more of these smoothly as possible. The more of these smoothly as possible. The more of these opportunities you take, the more of them opportunities you take, the more of them opportunities you take, the more of them you'll find and the more you find, the you'll find and the more you find, the you'll find and the more you find, the more you can do. And the more you'll more you can do. And the more you'll more you can do. And the more you'll find the limits of the tools that we're find the limits of the tools that we're find the limits of the tools that we're relying on now every day.
-
relying on now every day. relying on now every day. My last thoughts about where this is all My last thoughts about where this is all My last thoughts about where this is all going and more importantly, how we going and more importantly, how we going and more importantly, how we should feel about it. The amount of work should feel about it. The amount of work should feel about it. The amount of work somebody can do as a developer has gone somebody can do as a developer has gone somebody can do as a developer has gone up exponentially as a result of all of up exponentially as a result of all of up exponentially as a result of all of this. Any one dev can do 10 times to 100 this. Any one dev can do 10 times to 100 this. Any one dev can do 10 times to 100 times more work than they could before. times more work than they could before. times more work than they could before. They can't necessarily validate that They can't necessarily validate that They can't necessarily validate that that work is good or worth shipping, but that work is good or worth shipping, but that work is good or worth shipping, but they can absolutely get the code out they can absolutely get the code out they can absolutely get the code out more than they ever could. Your focus more than they ever could. Your focus more than they ever could. Your focus should not just be writing more code. should not just be writing more code. should not just be writing more code. Your focus should be solving more Your focus should be solving more Your focus should be solving more problems and you shouldn't be doing this problems and you shouldn't be doing this problems and you shouldn't be doing this out of fear, either. If you come into out of fear, either. If you come into out of fear, either. If you come into this scared that you're going to lose this scared that you're going to lose this scared that you're going to lose your job if you don't token max, you're your job if you don't token max, you're your job if you don't token max, you're going to burn a bunch of tokens and a going to burn a bunch of tokens and a going to burn a bunch of tokens and a bunch of your own sanity and come out bunch of your own sanity and come out bunch of your own sanity and come out way more stressed and not happy and way more stressed and not happy and way more stressed and not happy and probably still unemployed. If you go probably still unemployed. If you go probably still unemployed. If you go into this with excitement, the into this with excitement, the into this with excitement, the excitement that we all went into excitement that we all went into excitement that we all went into software with, with the ability to software with, with the ability to software with, with the ability to customize our computers and the things customize our computers and the things customize our computers and the things on them to do what we want and need on them to do what we want and need on them to do what we want and need specifically for ourselves. If you go specifically for ourselves. If you go specifically for ourselves. If you go into your agents with that mindset, if into your agents with that mindset, if into your agents with that mindset, if you go into the token maxing with that you go into the token maxing with that you go into the token maxing with that goal, trying to get more out of the goal, trying to get more out of the goal, trying to get more out of the things that you do and build every day, things that you do and build every day, things that you do and build every day, if you approach it with excitement, what if you approach it with excitement, what if you approach it with excitement, what you'll get back is incredible. When I you'll get back is incredible. When I you'll get back is incredible. When I went into this stuff skeptical, what I went into this stuff skeptical, what I went into this stuff skeptical, what I got out was some okay code and a bunch got out was some okay code and a bunch got out was some okay code and a bunch of bugs. When I went into these things of bugs. When I went into these things of bugs. When I went into these things excited to push the limits of what I excited to push the limits of what I excited to push the limits of what I could do and build, I came out with my could do and build, I came out with my could do and build, I came out with my own custom cloud. I came out with a pile own custom cloud. I came out with a pile own custom cloud. I came out with a pile of NPM packages I use every single day of NPM packages I use every single day of NPM packages I use every single day for a ton of different stuff. I started for a ton of different stuff. I started for a ton of different stuff. I started the forking the software I rely on to the forking the software I rely on to the forking the software I rely on to tailor it to my specific needs. I tailor it to my specific needs. I tailor it to my specific needs. I started building my own control planes started building my own control planes started building my own control planes for managing all of this.
-
for managing all of this. for managing all of this. I started having way more fun and I started having way more fun and I started having way more fun and building way more, too. So, don't building way more, too. So, don't building way more, too. So, don't approach this with the attitude of approach this with the attitude of approach this with the attitude of keeping your job and making sure you'll keeping your job and making sure you'll keeping your job and making sure you'll still be employed in a few years. still be employed in a few years. still be employed in a few years. Approach this with the excitement that Approach this with the excitement that Approach this with the excitement that we can build all the things we ever we can build all the things we ever we can build all the things we ever imagined. Lower your bar for what's imagined. Lower your bar for what's imagined. Lower your bar for what's worth building and raise your bar for worth building and raise your bar for worth building and raise your bar for how far you bother going. And what how far you bother going. And what how far you bother going. And what you'll find is a lot of fun, but also a you'll find is a lot of fun, but also a you'll find is a lot of fun, but also a lot of rate limits being hit. We're at lot of rate limits being hit. We're at lot of rate limits being hit. We're at 42% now and we still have 4 hours left. 42% now and we still have 4 hours left. 42% now and we still have 4 hours left. God. God. God. Yeah. Yeah. Yeah. I'm going to vote for my two subs, and I I'm going to vote for my two subs, and I I'm going to vote for my two subs, and I think you should, too. Mythos is an think you should, too. Mythos is an think you should, too. Mythos is an incredible model capable of things I incredible model capable of things I incredible model capable of things I never would have expected even a year never would have expected even a year never would have expected even a year ago for agents to be able to do. It ago for agents to be able to do. It ago for agents to be able to do. It builds awesome things. It writes awesome builds awesome things. It writes awesome builds awesome things. It writes awesome plans. It describes the stuff it's doing plans. It describes the stuff it's doing plans. It describes the stuff it's doing in ways that are actually digestible. in ways that are actually digestible. in ways that are actually digestible. And you can use this to push the limits And you can use this to push the limits And you can use this to push the limits of what your code is doing to get better of what your code is doing to get better of what your code is doing to get better information and spend less of your time information and spend less of your time information and spend less of your time doing the things you don't like and more doing the things you don't like and more doing the things you don't like and more of your time shipping things you do. If of your time shipping things you do. If of your time shipping things you do. If you find yourself stopping to ask, "Can you find yourself stopping to ask, "Can you find yourself stopping to ask, "Can an agent do this?" an agent do this?" an agent do this?" reset your mind. Just ask the agent to reset your mind. Just ask the agent to reset your mind. Just ask the agent to do it and see if it fails. And if it do it and see if it fails. And if it do it and see if it fails. And if it fails, try to figure out why. Ask it fails, try to figure out why. Ask it fails, try to figure out why. Ask it what it struggled with. Look through the what it struggled with. Look through the what it struggled with. Look through the history and the bad tool calls it made history and the bad tool calls it made history and the bad tool calls it made and try to steer it in the right and try to steer it in the right and try to steer it in the right direction. Maybe make a skill that direction. Maybe make a skill that direction. Maybe make a skill that points it in the right way. Maybe build points it in the right way. Maybe build points it in the right way. Maybe build your own verification in your code base your own verification in your code base your own verification in your code base to push it in the right direction. If to push it in the right direction. If to push it in the right direction. If you're not already writing your own you're not already writing your own you're not already writing your own custom lint rules for weird [ __ ] that custom lint rules for weird [ __ ] that custom lint rules for weird [ __ ] that your agents are doing, you're not being your agents are doing, you're not being your agents are doing, you're not being creative enough yet. And that's fine.
-
creative enough yet. And that's fine. creative enough yet. And that's fine. This is a different way of thinking, and This is a different way of thinking, and This is a different way of thinking, and it's a really exciting one. So, go out it's a really exciting one. So, go out it's a really exciting one. So, go out and build. Take advantage of these and build. Take advantage of these and build. Take advantage of these subscriptions, push the limits of the subscriptions, push the limits of the subscriptions, push the limits of the best models and tools available today, best models and tools available today, best models and tools available today, and don't get too locked into them and don't get too locked into them and don't get too locked into them because new things will be available because new things will be available because new things will be available tomorrow. Just try to be excited when tomorrow. Just try to be excited when tomorrow. Just try to be excited when you go in and you'll be amazed what you you go in and you'll be amazed what you you go in and you'll be amazed what you can pull out. Now that I got these can pull out. Now that I got these can pull out. Now that I got these agents queued, I'm going to go catch up agents queued, I'm going to go catch up agents queued, I'm going to go catch up on some sleep. I recommend you do the on some sleep. I recommend you do the on some sleep. I recommend you do the same, but come back fresh and build some same, but come back fresh and build some same, but come back fresh and build some cool stuff. Let me know what y'all are cool stuff. Let me know what y'all are cool stuff. Let me know what y'all are building, and until next time, building, and until next time, building, and until next time, peace nerds.
Summary
The main theme is the Fable AI model and its powerful, but costly, release on certain subscription tiers, which is being removed soon. Key subjects discussed include Fable (Mythos with safeguards), Claude Code, and the immense inference costs associated with its powerful capabilities. The practical takeaway is to take advantage of these generous, subsidized subscription plans for immense generative AI power within the limited time frame, even if a bit excessive for daily use.