Steven Pu retweeted
The coding interview is dead, at least for my team at Anthropic. I've seen enough to know that it no longer provides signal. I removed it from loops this month.
I have a spicy, optimistic take - I think there's an opportunity for junior engineers: felixrieseberg.com/hunger-be…
Steven Pu retweeted
We are announcing three new discoveries in mathematics, computer science, and physics about harmonic analysis, complexity theory, and quantum field theory.
openai.com/index/building-sc…
Steven Pu retweeted
An internal model at OpenAI has produced a claimed refutation of ETH! Russell Impagliazzo, Mohan Paturi, and I have been asked to help organize an open and independent examination of this claim. 1/4
Steven Pu retweeted
Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel.
- Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all.
- Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract.
- Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code.
- And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while.
The funnel is driven by a combination of factors:
1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel.
2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel.
3. Access. Free-tier, paid-tier, internal.
So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.
Judging by my tl there is a growing gap in understanding of AI capability.
The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code.
But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along.
So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions.
TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.
Steven Pu retweeted
I was talking with a friend today about how tough the job market is for tech, and how even if you do have a job, it's so unstable because things are changing so fast.
I told them that even though I technically have my own business and I'm supposed to have "visibility" into how I'm doing, do I really? No. I just assume the business will change every few months and I don't get too fussed when it does, and trust me it has been changing quite consistently
This is why it's obvious to me that, more than a set of skills or a knowledge base, you really have to teach your kids how to adapt to new situations, quickly figure out opportunities, and take risks
Steven Pu retweeted
We’ve lost an icon. Margaret Hamilton, a computer science pioneer best known for her work at MIT leading development of the flight software for the Apollo missions to the moon, has died at 90. Read about her legacy via MIT News: news.mit.edu/2026/margaret-h…
ALT Margaret Hamilton is reclined inside a NASA spacecraft module, surrounded by control panels and switches.
Steven Pu retweeted
Apple says you need a Mac to build iPhone apps.
Today I built one on Omarchy Linux, submitted it to the App Store, installed it on my iPhone over USB, and stopped it at a breakpoint in the debugger.
No Mac anywhere.
It's all open source:
github.com/joshuaswarren/oma…
Steven Pu retweeted
This is the doom I predicted a few days ago, coming for Photoshop. A clean-room open-source reimplementation.
No prizes for guessing that they decompiled Photoshop to source code, processed that to some kind of non-code specification language, then fed the spec to an LLM with an instruction to generate Rust.
Adobe just got nuked. And closed source is dead, dead, dead.
github.com/storytold/photocr…
Steven Pu retweeted
Let me tell you how I delayed WinXP SP2 for a month.
My software (BlackICE) was on the list of software that had to work. When they made major changes to the kernel network stack, my software broke.
They grabbed their reverse engineers to figure out why. What they discovered was that my software doing something evil, patching the kernel's network stack when our software was loaded. That's tricky on a multi-CPU system with warring instruction/data caches. We wrote a patent on how to do it reliably without crashing the system.
We searched through the kernel to find the location we needed to patch, looking for a certain pattern of bytes. They changed the network stack so that thing we were patching no longer existed.
So to make our software work, the inserted that pattern back into the kernel. As packets arrived, the new function would jump to that pattern, and then after, jump right back. This way, we could find the pattern and patch it -- and thereby patch the intended function.
Some Microsoft lawyers wanted to sue us for violating the software license agreement, which forbids reverse engineering, which we had to do in order to patch the kernel at runtime.
Cooler heads prevailed.
The point is our software did an ugly hack in order to work, and then Microsoft did an even crazier, uglier hack so that we'd keep working.
For years after, when meeting Microsoft engineers at conferences, they'd love to yell at me.
When Vista came along, they added signing of kernel models, putting an end to such nonsense, but also providing a real documented API we could use.
Microsoft once discovered SimCity was crashing on Windows 95 because SimCity had a memory bug.
Instead of saying “not our problem,” Microsoft added a compatibility workaround just for SimCity.
Because users don’t care whose bug it is. They care that software that worked yesterday still works today.
Backwards compatibility is a feature.
Steven Pu retweeted
Sure, but when it turns out the leak was caused by the plumber who neglected to tighten a pipe fitting, or perhaps poked a hole in one of the pipes, you shouldn't question the entire plumbing industry. You should question the plumber's compétence and/or motivations.
Steven Pu retweeted
If you showed me Claude Code today in 2018, I would have thought it was AGI.
We have already absorbed a dramatic amount of change in the profession of software engineering and society as a whole. But still, we’re starting to see cracks.
Things are accelerating faster than I can honestly stay on top of. Most people I know in AI are tired but powering through. I fear we will need more than that.
We need time to harden our systems, and for society to deliberate in how this technology is used and deployed.
I have a fairly low p(doom), I am confident that we can get through this! Humanity is incredibly reslient and adaptable, but this comes from our ability to make hard decisions- like this one, together.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must…
Steven Pu retweeted
Apparently every theorem can be proved just with linear algebra? Absolute W for representation theorists.
Steven Pu retweeted
This explains why I've been able to get so much done with vibe-coding while other people I keep recommending it to aren't quite able to have the same level of success: I have a strong CS foundation + verbal communication skills.
.... no pressure 😝
China published the most uncomfortable paper on vibe coding.
ETH Zurich tested 100 developers in a controlled, commercial-grade vibe coding environment to see who actually succeeds.
The findings are brutal.
The researchers tracked computer science achievement, written communication skills, and general cognitive reasoning.
They wanted to see what actually predicts vibe coding proficiency when you never touch a line of source code yourself.
Two major predictors emerged.
Written communication proficiency mattered. The ability to structure thoughts and articulate intent unambiguously in text directly impacts what the AI builds.
But that wasn't even the main takeaway.
Computer science achievement was a massive, dominant predictor of success.
Even when researchers controlled for general intelligence and reasoning skills, CS background still heavily dictated who built working software and who completely crashed.
In fact, CS knowledge contributed roughly twice the unique predictive variance of writing skills alone.
Why? Because vibe coding isn't about writing code. It’s about debugging logic.
When an AI agent builds a complex application and quietly breaks an edge case under the hood, a non-technical user looks at the glowing UI and assumes it works.
They don't know what questions to ask. They don't know what logic to challenge. They lack the mental models to recognize architectural catastrophe.
You can prompt your way past syntax.
You cannot prompt your way past a fundamental lack of engineering intuition.
The hype told us that learning to code is dead because language is all you need.
The data just proved the opposite.
To truly master the vibe, you still need to understand how the machine thinks.
Steven Pu retweeted
Dr. Dre Says the Only People Who See AI as a Threat Are Those 'Who Have Trouble Creating' variety.com/2026/music/news/…
Steven Pu retweeted
One of the reasons we keep working on a version you can run fully local.
And now the smaller models are getting good.
Steven Pu retweeted
They’re mostly trying to hit the impression threshold to get monetized.
This is why we don’t count replies anymore: it was the biggest incentive to spam.
Since we’re not paying them anymore, some of them have pivoted to selling paid promo slop to Polymarket, Higgsfield, etc.
Steven Pu retweeted
I find Opus 5 to be almost unusable.
It makes so many errors and has to be audited so often in order to ensure it doesn't drift off into laziness or being flat-out wrong that I think I'm done using it.
I'll get my Fable usage and then drop Claude for the week.
Steven Pu retweeted
It seems a common experience in the lives of people with high agency is that they, at some point, metaphorically, lean against a wall and discover that there is a hidden door. That some constraint they’ve been told about turns out not to hold, and this allows them to experience something positive. And from this they get into the habit of testing if walls are actually doors, and the more they do so, the more examples of hidden doors they find, and the more inclined they are to look for them. Far from the only thing that sets them apart, but having that experience can be transformative
— Henrik Olof Karlsson