Gabriel Lawson

@daniel_mac8 We're algomaxxed at the frontier now, until new advances. Its why OpenAI & Anthropic are productising so hard. Lock in that sweet customer loyalty to ward off against a competitor coming up with e.g. continual learning, new memory architecture or recursive self-improvement etc.

i love new models like gpt-5.5 as much as the next guy but when do we get new capabilities? last new LLM capability was o1 and reasoning since then, it's getting (much) better on the same dimensions guess the labs think new capabilities fall out of scale. Dario said as much. hope they're right.

@hunvreus But.. but.. scalability and... and resilience...

I really like Cloudflare, but I'm not gonna stitch together Workers + Queues + Durable Objects + R2 and Supabase on top of that just to do what I can do with a single deployment on Hetzner or Fly[dot]io. I can already see debugging is going to be a massive PITA.

@damianplayer Rewritten as a normal person would write it: If I give a clear instruction, managers must either tell me why it is wrong, ask what I mean, or carry it out. They are not allowed to sit on it, reinterpret it quietly, or ignore it.

this is why Elon’s companies move 10x faster than most. every founder should run their team like this: push back, ask, or execute.

The dust is settling with Opus 4.7. It's likely a new base model distilled from Mythos. The strategy seems to be dual-track: One model series for the haves (military, gov, powerful businesses) and a weaker distilled series for the have-nots.

@noahzweben @SebAaltonen Will you no longer be offering Team plans?! That will be devastating for some small businesses I'm advising.

@SebAaltonen We made this change like six months ago (back in November) due to enterprise customer demand for it. These pricing plans have been published on our site for months. Max plans are still the same

Anthropic removes all licensing options except the $20/month cheap license. Token priced after that. No discounts. I expect 2x-10x price hike for power users. Not surprising at all. Selling for loss wasn't good business. Others will follow eventually. letsdatascience.com/news/anthropic…

Claude Desktop helps you lock in with anthropic because you learn anthropic prescribed abstractions instead of actual agentic coding primitives. You're less likely to churn to OpenAI or others. This is why the app is being pushed so hard. Same story with codex.

@CurrenticAI Nope. The desktop apps for normie information workers not us.

This looks slick. Will you leave the terminal for this?

@AndrewCurran_ I think the reason anthropic did a restricted release for mythos is because it can autonomously chain zero day exploits across multiple attack surfaces. And not because they're worried about equity in society etc. That's an argument for caution rather than going full speed ahead.

There has been a great deal of speculation about why Anthropic is keeping Mythos in restricted release. One of the least-discussed reasons is cost. Not the cost to Anthropic of serving the model, but the downstream effects that cost will have on the industry, and on the world. Mythos is now being served to a small group of about 50 major companies. For organizations like these, token budgets are effectively unlimited, and the opportunity cost of not using as much of the model as possible is too high. I think you can already see the downstream effects even in this limited release. Claude users complain about hitting caps faster. They complain about degraded performance. For months now almost everyone I know has been continuously hitting the cap on Claude or Codex. The existence of Mythos pressures not just the amount of usage available to smaller subscribers, but also the pricing of these plans themselves, which are already subsidized. Smaller users will get hit twice. The compute cost of serving Mythos exerts pressure all the way down the line. Inference will get cheaper over time, but demand is already ahead of that curve and continues to expand. Mythos is not the end of this chain. As long as scale keeps rewarding larger runs, larger models will keep being trained. The next model that makes a Mythos-like jump may be dramatically larger again, and much more expensive to serve. If the cost of serving frontier models continues to outpace attempts to reduce it, then smaller players and public use get squeezed out. We end up with vast models, served at immense cost, available only to the richest corporations on earth. Those firms then use that access to outcompete smaller rivals, become richer still, and widen the gap again. If this continues, a small number of giant companies end up holding the only passports to the Country of Geniuses in a datacenter. For Anthropic, culturally, this is not a desirable world. Part of their reluctance to serve Mythos more broadly comes from a reluctance to help bring this world into being. There may be no way to serve a model like Mythos at scale right now without beginning this feedback loop. And as that loop accelerates, it will generate great resentment. If they serve it to lower-tier subscribers, those users get a handful of exchanges before hitting the cap. Seeing how capable the model is only deepens the resentment, because access is visibly rationed. The labs will be forced to make a trickle-down argument: let the largest firms use the models first, and the abundance will eventually spread to everyone else. The public is unlikely to buy this argument. The hostility and pushback against the industry will spiral. Eventually it may not remain merely political. It is not only Dario who has seen this world, but Sam as well. That is part of why OpenAI has started talking about mechanisms that would give ordinary citizens a direct stake in the upside of the industry, like the Public Wealth Fund. In my opinion the original use case of Worldcoin was a global UBI in a future where OpenAI won the race. Not only is that future no longer certain, but the trust and solidarity required to support a UBI no longer seem to me to exist in the West. The only path then is simply to scale everything as quickly as possible and hope abundance eventually arrives in a cascade strong enough that it reaches everyone on earth. To my friends who are in the safety camp, I understand this argument is hard to accept. Please consider that there is a level of capability beyond which, unless your p(doom) is literally 100, stopping becomes more dangerous than continuing. I think we passed that threshold even before Mythos. Even if stopping were possible - and I personally do not believe it has been for years - stopping here would lock in a dystopia. This dynamic is incentive-driven, just like the race itself, and just as hard to coordinate against. We must not stop inside this tunnel. The only way out is through.

@NovayaSaf @maddierune Wtf?

My children will grow up watching me feel things fully. Every uncomfortable emotion, every hard conversation, every moment I want to disappear and choose to stay present instead. That is the most radical thing I can model. A human being who does not run from being human.

@maddierune @ShadowofEzra What makes him scary to you? Genuine question from an Aussie..

@ShadowofEzra If I saw him in public I would have no idea who he was and quite frankly I’d be scared and avoid all contact

Dom Lucre responds to the backlash he has been receiving and says he is being treated unfairly because people are jealous of him being a celebrity, adding that he creates the news. Lucre says his advertising rates have been public for a while, and he is entitled to them because he gets 600 million impressions every month. “I am one of the biggest creators on X.”

@rohanvarma Now that I've rediscovered the cli, it would take a miracle to go back to any type of 'environment'. CLI makes it easy to switch between claude code, codex etc. That freedom from lock-in is worth it.

What would it take for you to stop using your IDE and completely switch to using the Codex App as your ADE (Agentic Development Environment)?

@trq212 /ultraplan would be nicer :)

/buddy

@jhon_hawk_5 @dkundel Any chance this plugin could be tested and supported on Windows as well? (I know I hate it too but stuck with it).

@dkundel omg! add support for review the plan mode! I was using a handcrafted method. 😂

I built a new plugin! You can now trigger Codex from Claude Code! Use the Codex plugin for Claude Code to delegate tasks to Codex or have Codex review your changes using your ChatGPT subscription. Start by installing the plugin: github.com/openai/codex-p…

@embirico This is clever and useful. Claude Code can do the grunt work while Codex provides the brains.

given the number of people who use codex to review claude, we decided to get a little cheecky

@tszzl Competence is downstream of so many things outside one's control that calling it a moral virtue basically just... lets the already competent feel 'righteous' about their contempt for those less fortunate.

malice and incompetence are so often confused because competence is an extremely important moral virtue most often missing from the world

@trq212 Interesting, thanks. The security subagent that Codex uses is indeed slow. Will try out this CC switch when I'm no longer living dangerously. :)

@glawsontweets that's not what this is, subagents are too expensive and take too long to run on permissions, we use a special classifier

@trq212 They call it guardian approvals; When Codex needs approval for higher-risk actions (e.g. blocked network access), it routes eligible approval requests to a security reviewer subagent rather than blocking on human input. I thought you guys built it as feature parity!

@glawsontweets no one has this :)

@polynoamial @saranormous @karpathy @NoPriorsPod He's escaped the dark side and no longer has to answer to.. those who pay the compute bill. Once you've tasted that you tend to not want to go back.

@saranormous @karpathy @NoPriorsPod Why is he not at a frontier AI lab at the most pivotal time in human history since at least the industrial revolution?

excited to have sensei @karpathy back on @nopriorspod tomorrow. what wisdom shall we glean? (yes, yes, wth is going on with coding agents)

@bcherny /btw works nicely but its like the movie Memento. every /btw chat turn has no idea about previous /btw chat turns. Can this be... helped, may be?

btw