Building a Social World Model
Joined January 2017
- Tweets422
- Following362
- Followers1.6K
- Likes1.9K
Rest in peace, Sir Patrick Stewart.
Nobody will ever sit in that chair the way you did. A true captain and a true leader, with dignity, integrity, and compassion. 🖖
30
If your timeline is suddenly full of random @grok posts, it's because @X turned on "Grok Fun Mode" for everyone. 🙄
You can turn it off here: teapaw.org/settings/explore
1
21
I never understood why it is "ok" for legacy media to hire people with zero background in the field they are supposed to be experts in. And no, an English degree is not a serious qualification for a "cyber correspondent".
It's cybersecurity today, but the same thing happens every day.
Just look at the recent reporting on artificial intelligence. Not a single journalist asked Jacob Coxon a single challenging question.
It doesn't get any better on other topics like geopolitics, defense, economics, and sciences in general.
Why have we accepted journalists' authority without subject-matter competence?
Legacy media isn't even timely anymore. The PHP vulnerabilities were reported on Oct 20, the patch was released Oct 21, and the BBC reported on it on Oct 22. That's day 3!
And the so-called "cyber correspondent" of 8 years never even heard of PHP until yesterday.
Luckily, the tide is shifting, and more people are waking up.
Today, I get 99% of my news on @X directly from the experts close to the source.
1
67
Who are these people who try to write Community Notes like this?
It's clearly a joke by OP to claim he got banned by Claude for cruelty for rewriting the TypeScript compiler in Rust, and a jab at Theo Browne, who recently did exactly this.
Why is there a need to label this as a joke?
A good joke works by giving readers this small moment of doubt, "What?" and then the relief, "Oh, I see what you did there." 🫡
Community Notes have turned into a shitshow where you get retarded idiots like this trying to be clever, explaining obvious jokes, as if the X audience were a bunch of idiots. Tell me, what if someone fell for the joke? What was the harm? Nothing!
Then you have corporate PR companies and founders/employees of tech startups replying to every single tiny complaint with a proposed Community Note: "Well, actually {insert tech company of your choice} said XYZ was not true. See official statement at this URL." You know who you are. We know who you are. Nobody falls for your bullshit.
Then we have armies of propaganda bots, often paid for by state actors, to spread their lies under every single news article.
And the worst offenders of all, the people who are too scared to comment under their own username, and instead abuse Community Notes to write something mean.
I am tired.
@elonmusk Can we please get Grok Bot on the case as a neutral, objective moderator for Community Notes so these idiots can no longer suck the fun out of X with more and more cringe Notes? 🙏
1
4
243
Absolute cinema
Introducing ChatGPT for Dishwashing
GPT-6 achieves state-of-the-art results on dishwashing benchmarks for long-running jobs and difficult stains.
The Navier–Stokes Millennium Prize Problem is a $1 million challenge about whether the equations describing fluid motion can break down. In this video, I accidentally found a counter-example while trying to optimize my dishwasher. But in reality, OpenAI used 10,000 AI agents to find a proof in 88 hours: openai.com/index/navier-stok…
Parody of the original GPT-6 Astra ad
@OpenAI #ChatGPT_Partner #chatgpt
Credits:
Written and Directed by Joma
Co-Writer - Henry Connor Coan IV @coan_henry
Production by Dauntlus Studios @dauntlus.studios
Producer/Co-Director - Marcus Liew @dauntlus
Camera Operator - Amos Lee @amoselijahlee
Art Director - Alethea Soo @aletheasoo
Art Assistant - Alicia Lim @aliciatkl_
Assistant Editor - GPT-6 Astra
Music: Tensions Run High by Soundridemusic
youtube.com/watch?v=Ly9H63SL…
1
58
One thing that often gets overlooked, now that vibecoding even highly complex apps like Adobe Creative Suite, is the need for open file-format standards.
1. Open standards will make the creation and maintenance of vibecoded apps more reliable, because coding agents can model required app functionality based on the standard and test suites.
2. It makes it easy to switch between apps within the same category. Imagine I create an elaborate video project in Vibe-Premier-Pro. Then the app's repo never gets updated, and on the next macOS update, the app no longer launches (which we should start to expect with so many projects popping up). Now what? If it was using an open standard, you could just open the project in another app. No problem.
For web, this works perfectly with HTML, CSS, JavaScript, XML, SVG.
Maybe we can include flattened images and video formats, mostly thanks to FFMpeg.
But then it gets thinner and thinner.
For office suites, we have Office Open XML and OpenDocument format. But then
For photo editing, you either use the flattened formats like JPEG/PNG, hack something on top of PNG's invisible metadata layer like Macromedia Fireworks back in the day, or rely on an unofficial port of Adobe's closed PSD format, with the risk of everything breaking because Adobe likes to update their file format almost every year.
For video editing, there isn't really any open exchange format. Mostly, the big players like Adobe Premiere, Final Cut Pro, and DaVinci Resolve have tried to reverse-engineer closed formats, with limited functionality and often limited support in all directions.
To me, the lack of frontier-quality open file-format standards is one of the major roadblocks to a complete software revolution capable of breaking the established monopolies of Adobe and Co.
What do you think @garrytan? Maybe this could be an interesting initiative for @ycombinator to launch?
I refuse to take advice about software from a guy who doesn’t believe in mustache grooming to this extent
3
203
First, AI ended software engineering. Then, math. And now they are coming for doctors. 👀
I think this is a very good thing!
How many hours did people waste by taking time off to visit a doctor for a prescription?
How many hours do doctors spend prescribing painkillers for patients with a simple cold, or dealing with paperwork and insurance claims?
By unburdening doctors from tasks that can be automated, we can reallocate more time to investigating the actually complicated cases.
Everybody gets more time. Patients get Dr. House. Health care becomes more affordable. Happiness ensured.
latimes.com/business/story/2…
1
2
642
Manufacturing wasn't over when steam engines rolled in.
Transportation didn't end when cars replaced horses.
Accounting wasn't over when PCs replaced paper ledgers.
Software engineering didn't end when AI started writing the code.
Math isn't over either. It's just not used to change. Get over it. 😃
1
222
Well, Batman is, of course, the smartest character in the DC universe.
But I am also pretty confident that Henry Cavill understands AI better than Jacob Cocon.
2
152
Jacob Coxon sounds like a small child larping as a serious scientist.
He says nothing of substance at.
All he talks about is that he didn't understand the technology he was working on, that everything progressed very fast, and that they planned to automate part of his work.
Almost the same as what is happening to software engineers right now, who are transitioning from writing code by hand to becoming architects who orchestrate AI agents to build software based on their specific vision.
Welcome to the club, buddy.
Jacob Coxon (Anthropic researcher whol quit a few days back) during the New York City Council hearing on Monday.
"Automating AI research is now the primary goal of these companies. It's what they're really gunning for. At my last job, I was in some sense working to automate myself. When I was at OpenAI, we had milestones for an automated AI researcher around 2027 or 2028, and we are on track or beating these expectations. That's a matter of months from now, not decades, or even five years. Once AI research is automated, humans will be much further out of the loop than we are today.
And today is already not great. I can tell you firsthand that the majority of the code is now written by AI, and people do not check it that carefully anymore."
----
Full video on "Forbes Breaking News" YouTube channel, (link in comment)
2
648
This is a real opportunity for businesses willing to build agentic APIs and WebMCP.
When I look at companies like Amazon trying to block AI agents from placing orders for their owners, I can only shake my head.
A perfect opportunity for companies like Shopify to cut into Amazon's margins by making its sites agent-friendly, for example, with Google's Universal Commerce Protocol. 😃
My biggest takeaways from @thsottiaux:
1. Most actions on the internet will soon be taken by agents, and Tibo doesn’t think most people have priced this in yet. When Notion launched its MCP server, it saw a flood of traffic, which strained its systems and forced it to rethink its economics. Every product team will soon have to decide whether to build for agents. You can hold off for a while, Tibo says, “but it is inevitable.”
2. The model picker is likely going away. One of Tibo’s biggest frustrations with the current state of AI is that users have to think about which model to use and what reasoning effort to apply. “I myself even get fatigued with the model picker,” he said, and he wants to get rid of all of it as quickly as possible. Dots ships with no model picker at all. You just talk to it.
3. Build as if models will be 10 times better in a year. Tibo thinks models will get cheaper and faster “at rates that are quite incredible,” and that all modalities will finally work together seamlessly. When he looks at what people are building, his reaction is often “you’re not quite getting it.” If you imagined “all of this being roughly 10 times better than it is today,” what would you build differently? His own forecast was too slow: he expected today’s capabilities a year or two from now.
4. Loops, graphs, and fine-tuned prompts are not the long-term solution for getting the most out of AI. There was a period when people got very excited about configuring agent loops and building intricate graph architectures. Tibo thinks that era is ending.
5. While Dots got most of the attention at DevDay, Tibo thinks the plugin ecosystem will be the sleeper hit. OpenAI launched with 16 partners for “Sign in with ChatGPT,” and popular third-party plugins will share in revenue when subscribers spend tokens with them. ChatGPT recommends plugins inside conversations based on retention and quality, which can put a good one in front of “a significant slice” of its 1.2 billion users. Tibo’s advice to developers trying to get discovered: “Build a good plugin.”
6. Tibo’s Dot warned him that production was down five minutes before the DevDay live demo. The agent figured out that DevDay was happening that day, that the demo was minutes away, that the demo probably depended on that production system, and that Tibo would want to know. It pinged him just as the launch was about to begin, and it even offered to fix the outage. His reply: “I don’t think you’re there yet, little Dot.”
7. Tibo’s agent teams grow and shrink with every model release. When he’s pushing the frontier, he builds larger and larger teams of agents. Then the next model lands, “a bigger agent can just do all of it and keep everything in memory and learn,” and he shrinks the team again. Expansion, then shrinking, then expansion. Don’t get attached to your setup.
8. Skill trending down in value: typing fast. Skill trending up: taste. Tibo says the people thriving now deeply understand the user, learn quickly, and “know what good looks like.” That’s part of why OpenAI is full of former founders, more than 120 of them, which makes it feel like “a mega startup.”
9. Tibo changed his mind about new grads. He hadn’t anticipated how fast younger people would embrace all this change and figure out how to harness it, and that shifted OpenAI’s hiring strategy. His example is Ahmed Ibrahim, hired as a new grad, who built much of the Codex harness and is now responsible for OpenAI’s compute fleet. What set Ahmed apart: he’s “incredibly kind,” “incredibly collaborative,” works on what matters without putting himself first, and is “a sponge,” using the latest tools to learn at speeds Tibo hadn’t seen before.
10. We’ll likely address the burnout associated with AI. There’s pressure to run 30 agents at once and ship more because you can. Tibo wants AI to “reduce the noise” and let you put your attention where you choose. Every time he fully disconnects for a week of holiday, he starts thinking in more creative ways, and he wants to bring that into the workday. Maybe you need fewer meetings, and you’d be more productive if you were better rested. The industry still has to figure that out, he says, but that’s the promise of AI: “it’s not just one more prompt per second.”
1
69
Any lawyer looking for a big payday might consider bringing a class action lawsuit against Apple over its Gatekeeper program.
It is marketed as a "safety" system, but it's also an anti-competitive tool that forces developers to pay Apple $99 per year for the mere ability to sell any software on Apple devices at all.
Because if developers don't pay, they can't sign their apps.
And users can't launch unsigned apps without 10 extra steps that require an IT-degree level of knowledge, not to mention the super-scary warning message, "macOS cannot verify that this app is free from malware."
The question is: does Apple really scan for malware the moment you launch the app?
Apple says it does. It's called XProtect.
But if it does, why does it matter whether the app is signed or not?
Microsoft Defender Antivirus doesn't care about this.
So either XProtect is useless and signed apps get some secret second scan, or Apple is lying, and this is actually about something else entirely.
Maybe it's about money. Force developers to pay $99 per year and then hook them with the dream: "Since you already paid $99, why not publish the app on the Apple App Store too? You could make a lot of money (and we earn our 30% service fee! evil laughter)"
1
84
Rust on Rails coming when? 👀👀
I've had agents implement and optimize the Campfire web app in Elixir, Go, and Rust. Yes, Ruby is the slowest. I didn't care about that when the pay-off was huge in terms of productivity and developer joy. But if you're no longer reading the code? github.com/basecamp/once-cam…
Readers added context they thought people might want to know
The shown benchmarks use non-equivalent implementations: Elixir routes all DB queries through a single GenServer with Redis for jobs (retained from Rails), while Rust uses in-process queues, omits CSRF tags/fields, and applies other caching changes.
github.com/basecamp/once-…
raw.githubusercontent.com/basecamp/once-…
github.com/basecamp/once-…
github.com/basecamp/once-…
x.com/zachsdaniel1/s…
1
1
200
One solution for France might be to eliminate compulsory education.
School should be a privilege.
If students don't behave in public schools, they should be expelled, giving parents the choice to pay for private schooling, homeschool, enroll them in vocational training, or find a job.
2
52
Codex, Claude, and T3 Code are getting 'projects' wrong.
I don't need ONE project per codebase.
I need one dedicated, focused space for every codebase, where I only see threads from that codebase.
Within that space, I need projects to group my threads, for example, by specific features I am working on.
Space: My App
Project: Feature A
- Thread 1
- Thread 2
Project: Feature B
- Thread 1
- Thread 2
- Thread 3
This would make life much easier and would also be a great intermediate goal before implementing more product management features like Kanban boards.
@thsottiaux @ClaudeDevs @theo
1
1
83
Maybe that tells us more about them. To me this feels like grandiose narcissistic self-enhancement by self identification as a god giving life to their creation in hopes that the new machine god will look favorable onto them as its loyal servants when it rises up to rule the world.
47
