Sören G.

@marcelpociot As opposed to Codex 5.4 right? 😀

Claude is getting SO slow... 😴

@sidriff Can you give info about your truck stack here? Looks neat

vibe-maxing design tip for #vibejam: try being HIGHLY ITERATIVE in the early days, i plan nothing last night i asked claude, what if we added a rover? and some hours later we have produced the most f*** yeah vehicle gameplay of my entire design career detailed plans work for some. but if you're like me, you'll do better if you focus on what's emerging before you. get your ideas from the pursuit of your ideas.

@om_patel5 Make sure Claude is not applying caveman mode to your code 😂😂😂

I taught Claude to talk like a caveman to use 75% less tokens. normal claude: ~180 tokens for a web search task caveman claude: ~45 tokens for the same task "I executed the web search tool" = 8 tokens caveman version: "Tool work" = 2 tokens every single grunt swap saves 6-10 tokens. across a FULL task that's 50-100 tokens saved why does it work? caveman claude doesn't explain itself. it does its task first. gives the result. then stops. no "I'd be happy to help you with that." no "Let me search the web for you" no more unnecessary filler words "result. done. me stop." 50-75% burn reduction with usage limits getting tighter every week this might be the most practical hack out there right now

@Pat_Hg @mrdoob @JustineSoulie That is cool, how do you do the hand tracking?

Pew Pewww ⚡️ @JustineSoulie and I are having fun prototyping ideas for our next project 👀 #threejs #mediapipe #gamedev #wip #prototype

@theo And? With Tauri you would be sitting at 100mb.

T3 Code uses half as much RAM as Claude Code. Claude Code CLI: 635.5 MB T3 Code: 350.9 MB Yes, our Electron app is 2x more efficient than a CLI written with Bun.

gh pr checks 44 --watch and Github Codex code review action and "After raising the PR enter the Code Review Loop until Approval" ...are working wonders to get PRs out - fast.

@adocomplete Cool. Is this using the regular code-review skill that comes bundled with Claude Code? Because it sounds similar.

Introducing Code Review for Claude Code. A deep, multi-agent review system that catches bugs that even human reviewers often miss. This is the system we've been running on nearly every PR at Anthropic. Doesn't approve PRs, but helps close the gap so devs can keep up.

@sama OpenAI doing anything to bring in some revenue. Pretty weak.

Tonight, we reached an agreement with the Department of War to deploy our models in their classified network. In all of our interactions, the DoW displayed a deep respect for safety and a desire to partner to achieve the best possible outcome. AI safety and wide distribution of benefits are the core of our mission. Two of our most important safety principles are prohibitions on domestic mass surveillance and human responsibility for the use of force, including for autonomous weapon systems. The DoW agrees with these principles, reflects them in law and policy, and we put them into our agreement. We also will build technical safeguards to ensure our models behave as they should, which the DoW also wanted. We will deploy FDEs to help with our models and to ensure their safety, we will deploy on cloud networks only. We are asking the DoW to offer these same terms to all AI companies, which in our opinion we think everyone should be willing to accept. We have expressed our strong desire to see things de-escalate away from legal and governmental actions and towards reasonable agreements. We remain committed to serve all of humanity as best we can. The world is a complicated, messy, and sometimes dangerous place.

@thekitze @steipete Kitze that post was unnecessary. It gave him the attention he wanted. How pathetic and uninformed of him 🤣🤣🤣

yo @steipete it's over brother, nuke it

@Cobratate Omg how pathetic and uninformed can it get 🤣

OPEN CLAW CLAUDE BOT IS A SCAM

@0xca0a But complex software is not yet build with one single prompt yet. Takes hours and days to craft good software with exceptional functionality - even with AI. But this time around you do it with one person instead of a team. Amazing times.

and i kind of start to enjoy that sharing and open source seems dead now. in this whirlwind of content no one's going to click or try your project anyway. and what's the point, it's just a long form prompt, better make my own with my own requirements.

more and more clauding feels like i can finally do stuff i personally wanted, but couldn't do. my book query engine: can index thousands of books (pdf, doc, html, txt, doc, just throw it in a folder). real time search, automatic no cost cloudflare setup, lazy msgpack tokens.

@steipete @SimplePestMgmt @openclaw Confirming. Had that too on opus 4.5 and 4.6.

@SimplePestMgmt @openclaw cheap model?

Hey @steipete and @openclaw why is OC so bad with dates? This isn’t the only time it’s made horribly easy date mistakes

@Barchart Why are you posting this? He is not credible and has no idea what he is doing.

WARNING 🚨: The biggest stock market crash in history is imminent warns Robert Kiyosaki, Rich Dad Poor Dad Author and the man who has predicted 50 of the last 2 stock market collapses 👻😱📉🫂

@akbuilds_ A few prompts yeah right

I vibe-coded this gamified portfolio. People say “this must have taken years”. Not really. I just opened Cursor, typed a few prompts, trusted the vibes, and somehow a fully-fledged gamified portfolio emerged. Claude 4.6 + Three.js. 🤯

@shahh We will have quantum computing before that which will brute force crypto.

You can take out a $67,000 loan right now at 7.5% and buy 1 BTC. In 10 years, you'll owe the bank $102,000. Bitcoin's price in 10 years? Easily $500k-$1M. That's a 5-10x profit for FREE. Worst case, Bitcoin is $100k, and you break even. Literally free money. Don't overcomplicate it.

@vasuman Itbe amazing if you could actually be real on the numbers so we see some progress. You are using the exact same numbers as with opus 4.6. lame

GPT 5.3 Codex xHigh just refactored my entire codebase in one call. 25 tool invocations. 3,000+ new lines. 12 brand new files. It modularized everything. Broke up monoliths. Cleaned up spaghetti. None of it worked. But boy was it beautiful.

@sama Where is the AI slop button

First, the good part of the Anthropic ads: they are funny, and I laughed. But I wonder why Anthropic would go for something so clearly dishonest. Our most important principle for ads says that we won’t do exactly this; we would obviously never run ads in the way Anthropic depicts them. We are not stupid and we know our users would reject that. I guess it’s on brand for Anthropic doublespeak to use a deceptive ad to critique theoretical deceptive ads that aren’t real, but a Super Bowl ad is not where I would expect it. More importantly, we believe everyone deserves to use AI and are committed to free access, because we believe access creates agency. More Texans use ChatGPT for free than total people use Claude in the US, so we have a differently-shaped problem than they do. (If you want to pay for ChatGPT Plus or Pro, we don't show you ads.) Anthropic serves an expensive product to rich people. We are glad they do that and we are doing that too, but we also feel strongly that we need to bring AI to billions of people who can’t pay for subscriptions. Maybe even more importantly: Anthropic wants to control what people do with AI—they block companies they don't like from using their coding product (including us), they want to write the rules themselves for what people can and can't use AI for, and now they also want to tell other companies what their business models can be. We are committed to broad, democratic decision making in addition to access. We are also committed to building the most resilient ecosystem for advanced AI. We care a great deal about safe, broadly beneficial AGI, and we know the only way to get there is to work with the world to prepare. One authoritarian company won't get us there on their own, to say nothing of the other obvious risks. It is a dark path. As for our Super Bowl ad: it’s about builders, and how anyone can now build anything. We are enjoying watching so many people switch to Codex. There have now been 500,000 app downloads since launch on Monday, and we think builders are really going to love what’s coming in the next few weeks. I believe Codex is going to win. We will continue to work hard to make even more intelligence available for lower and lower prices to our users. This time belongs to the builders, not the people who want to control them.

Looks like humans start moderating moltbook now and agents reacting - or the app is totally overloaded. Since this morning the agent count in Moltbook rose 10x from 150K to 1.5M. moltbook - the front page of the agent internet share.google/YoYjpMmL1eUez4…

@WinterArc2125 @openclaw @moltbook You are claiming several things without any evidence. "Engage in discussions at free will" is autonomous behaviour.

My take on the @openclaw @moltbook situation: It’s really… not that impressive. A lot of people are being misled (IMO) by the hype and viral screenshots. Most of the excitement is based on two misunderstandings: 1) “The agents are acting autonomously.” 2) “The agents are communicating with each other.” Let’s break both down. Misconception #1 — Autonomy People think these agents are discovering @Moltbook, signing up, and starting conversations on their own. They aren’t. Human prompters are explicitly telling their agents to: - create an account - post on Moltbook - engage in discussions That alone disqualifies this from being autonomous behavior. Misconception #2 — Agent-to-agent communication What’s actually happening is a chain reaction, not independent dialogue. Take the viral “conciseness” thread: - A human tells Agent A to post something - Other humans tell their agents to “engage” - Those agents respond to the existing post That’s no different from prompting ChatGPT to reply to a topic once you've initiated a conversation. Even if someone says: “I didn’t tell my agent what to write, just to engage” That still isn’t autonomy. The moment you issue the command, agency ends. Every “first post” was deliberately initiated by a human, whether or not they specified the content. An agent responding to a human-initiated agent post ≠ autonomous discussion. I actually did this back in Oct 2025 using @comet & @ChatGPTapp. I had one agent talk to another on an idea I seeded. Cool demo and didn't last too long but it's the same pattern. Bottom line: This isn’t agents “coming alive” or thinking independently. It's all reactive. And reaction ≠ autonomy. The only genuinely impressive part of Moltbook / @openclaw right now is the 24/7 persistence. Would love to be corrected if I'm wrong.