Muzoweri

Retweeting interesting stuff into the void. Mostly AI coding/cool tech & science. Occasional reply guy. Normal disclaimers apply.

[Lenny Rachitsky @lennysan]

My biggest takeaways from OpenAI's Codex lead @ajambrosino:

1. Product work has inverted. The old product process was built around the assumption that building things is expensive, so de-risk everything up front with specs, research, and prototypes. That assumption is gone. The hard work has shifted from “Should we build this?” to “Of all the prototyped attempts at this idea, what's the best idea, what should we fold together, and what do we go all-in on?”

2. Your role is now defined by the average of what you spend time on. Deginers write code, engineers do design, PMs ship. So what are you? You're now defined not by your title but by how you spend your time. If you averaged out everything you do in a week, where do most of those dots land? That’s your role.

3. Codex PMs use a "zone defense" strategy to stay on top of everything. With ideas flying at them from every direction, top-down annual planning doesn't work, so they spread their team out to cover the whole company. If two product people are working too closely, without any gaps, that's a bad sign. They space out PMs across the org for full coverage, and backfill gaps with product-minded engineers.

4. What is AI so bad at design? For two reasons: one practical, and one structural. Practically, design is harder to grade than code, and labs prioritize coding because it accelerates AI research. Structurally, good design requires novelty and culture—a model that outputs the @Linear website every time isn’t showing taste—and there’s a visual-to-code abstraction layer models can’t yet bridge. The practical reasons will likely be solved; some deeper challenges around novelty, culture, and abstraction may persist.

5. The original Codex Web release was “too AGI-pilled for the moment.” The first public Codex release was built on too ambitious a premise: give the model a task, and it comes back with the task finished. The problem was that the models at the time weren’t good enough to deliver on that promise reliably. Claude Code launched locally, asked questions, and sat with the user—a much better fit for where model capability actually was. Andrew thinks about that this constantly: are we building for where the models are, or for where we wish they were?

6. Andrew is confident that the Codex app launched in February 2026 would have failed if it had shipped in November 2025. The product was identical—the models were not. The lesson he learned was to keep prototypes that aren’t ready yet, and revisit them with each new model generation. Resist the temptation to kill a feature just because the experience isn't perfect. “It might not be ready yet” is very different from “it’s a bad feature.”

7. Taste isn’t just about aesthetics—it’s deciding what to build when you can build anything. Andrew points to a tweet arguing that people overemphasize taste’s aesthetic side (the example: Paul Graham has great taste and wears cargo shorts). Real taste blends aesthetics with systems thinking: knowing the direction, the theme, and how to present an idea. Ask “If we can build anything, what should this be?”—which he says is now the most important decision to make, in every field.

8. The design process isn’t dead. Yes, the formal design process as taught in design schools is finished. What remains is the meta-awareness of where in the product development process you actually are. The danger Andrew sees is the fully polished prototype that looks production-ready before anyone has done the research, and a roomful of people who assume it’s further along than it is. “That’s the design process now,” he says, “multiplayer exploration that looks like a finished product.”

9. “PRDs are dead” is also completely wrong. Because implementation has become cheap across every format, it’s tempting for non-engineers to jump straight to prototypes and for engineers to write long documents—when neither is the right tool. Andrew’s rule: if you’re trying to establish product clarity around a vague area, it’s probably a document; if you’re stress-testing an interaction pattern, it’s a prototype. The medium used to carry an implicit signal about where you were in the process, and now it doesn’t.

10. Most careers are longer than any one moment of failure. Andrew’s current success at OpenAI is, in his telling, 10 to 15 years of accumulation: skill set, passion, and market timing finally lining up at once.

[Muzoweri @muzoweri]

@RichardTyre8 @JamesMelville He's not really angry. It's exaggeration for comic effect for the vid. He's just explaining the rules of the road, not his private theory.

[David Pan @davep]

Everyone is writing about agent loops right now. Including us at Cursor, because they're so powerful. But here's a prediction: a year from now, nobody will be talking about them.

Not because they weren't useful. Because they'll work right out of the box. Batteries included. No instructions necessary.

Feels a lot like prompt engineering two years ago. It was incredibly important. People wrote courses on it. Now you just talk to your agent like a normal person.

That's the strange thing about AI right now. You learn something critical, get huge gains, and before long it's the new normal and something else is the bottleneck. So the alpha isn't what you know. It's how fast you learn it, and how easily you can let it go.

[tianhang zhu @TianhangZhuzth]

Hey guys, I'm leading LLM research for fundamental research labs. I'll be posting more.

[6529 @punk6529]

"The fat was always the point. The salad was just keeping it company."

I don't have a good word for the style (example above) of this fake-profound LLM last sentence in an essay structure, but it is driving me crazy.

It is everywhere right now. 100x worse than em dashes

[Theo - t3.gg @theo]

Om Malik passed away today. He was one of my favorite writers. His pieces on the chaos and self-inflicted pains of Silicon Valley were truly novel. I will miss hearing his thoughts.

If you haven't read this New Yorker piece, I highly recommend it. Lots we can still learn from.

[Muzoweri @muzoweri]

@shashj @infantrydort To save anybody else drilling into this @infantrydort had an anti-foreigner brat attack against Brit @shashj. It was rude but not a "racial slur". @shashj you will alienate people willing to listen if you pull that shit.

[Muzoweri @muzoweri]

@PaulEmbery "Faithful Labour MPs unite loyally behind man in specs"

[Ethan Mollick @emollick]

A fundamental problem with extending Codex/Cowork/Code to all knowledge work is that they remain very "software-brained" where the end result (the software) is what is important & that code serves as a source of truth.

For a lot of other knowledge work, the process is at least as important as the outcome. This includes researching what is known, an exploration of alternatives, failed efforts, prototype branches, experiments, etc. All of those things are valuable, so you cannot use the PowerPoint at the end the way you can use a codebase, nor is progress on a to-do list sufficient context post compaction. You work in learning loops, refining your perspectives as you go.

In some ways, this makes long-running models like Fable hard to use for deep knowledge work, since they are designed to deliver product to you in the end. You can prompt your way around this problem, but everything about the Codex and Code harnesses want you to be a software developer and you have to fight them. There is a real disconnect between how a manager or analyst thinks about problems and how the agentic software tools approach solving them. Addressing this is critical to breaking out of the coding niche for these tools.

[Matt Van Horn @mvanhorn]

https://t.co/E7cn84qhZk

[signüll @signulll]

cold open: google campus. a conference room named “moonshot serenity 4b.” twelve people are in a meeting titled: pre-sync for sync alignment on ai velocity.

sundar sits calmly at the head of the table.

a pm clicks to slide 1 of 187.

“the agenda today is simple,” she says. “how do we move faster while preserving our culture of not doing that?”

everyone nods.

then the door opens.

noam shazeer walks in.

the room goes silent.

noam: “i’m leaving.”

a vp of gemini reliability, brand, trust, latency, policy, and vibe raises a hand.

“leaving… this meeting?”

noam: “google.”

someone gasps. someone else opens a doc titled retention narrative draft final final noam v7.

sundar blinks once.

“noam, we brought you back.”

“for two point seven billion dollars.”

“technically you licensed some technology and reacquired talent.”

“that sentence is why we need legal in the room.”

legal is already there.

cut to: openai.

sam altman stands beside a whiteboard that just says ship.

an engineer walks by carrying a server rack and what appears to be the future.

sam: “we can offer speed, compute, and one meeting.”

noam: “one meeting per week?”

sam: “no. one meeting. total.”

back at google, the emergency retention committee forms instantly. it has 31 members.

a director says, “what if we give him a new title?”

“he already co-leads gemini.”

“distinguished super co-lead?”

“google fellow?”

“he already left google, founded a company, got brought back for billions, then left again. he’s folklore.”

meanwhile, a gemini launch review begins.

pm: “we’re ready to announce the model.”

policy: “can it answer questions?”

eng: “yes.”

policy: “too risky.”

marketing: “can we call it experimental?”

research: “the model is better than the last one.”

brand: “better is aggressive.”

trust & safety: “what about ‘more contextually adjacent to usefulness’?”

a staff engineer whispers, “openai just shipped a model while we were discussing the adjective.”

cut to noam’s exit interview.

hr: “what could google have done better?”

flashback montage:

a chatbot blocked because it might be too good.
a launch delayed because a button was the wrong shade of responsible blue.
a spreadsheet comparing twelve ai product names.
a meeting where someone says “we need a single coherent ai strategy” and three new strategies are created before lunch.

noam: “nothing comes to mind.”

hr: “great. we’ll mark that as positive attrition.”

later, sundar calls him privately.

“google is still google. best researchers. best infrastructure. billions of users.”

“yes.”

“so why leave?”

noam looks out the window.

“because you have everything except permission.”

silence.

sundar, softly: “we can create a permission working group.”

cut to all-hands.

sundar addresses the company.

“noam is leaving. this is not a loss. it is an opportunity to reflect on our operating model.”

chat explodes:

“is this recorded?”
“which gemini?”
“can we ask gemini why people keep leaving?”
“it said ‘insufficient context.’”

a vp steps up.

“to honor noam’s legacy, we’re launching project attention.”

applause.

“it will study whether attention is, in fact, all we need.”

a researcher raises a hand. “didn’t we answer that in 2017?”

“yes. but now we need enterprise readiness.”

final scene: noam arrives at openai. badge works instantly.

receptionist: “yeah, we just made one.”

no pre-read. no doc. just a whiteboard, five people, and a model running somewhere hot enough to toast bread.

sam: “ready?”

noam smiles.

cut back to google. a calendar invite appears:

meeting: reduce meetings task force kickoff
duration: 90 minutes
required attendees: 214

sundar sighs, opens gemini, and types:

“how do we move faster?”

gemini responds:

“have you considered leaving google?”

smash cut to credits.

[@levelsio @levelsio]

I don't know if it's obvious information or not but if you talk to random people in San Francisco the general thing they say is that software is commoditized cause so easy to make anything with AI fast (like how I cancelled all my SaaS subscriptions and just vibe coded a replacement for free) and that everyone smart is getting into hardware cause it's still difficult to enter, kinda related to the Midjourney Medical thing too

[Muzoweri @muzoweri]

@JebraFaushay More likely they will be watching DiscoFoot

https://t.co/oxUjgD9JiL

[Muzoweri @muzoweri]

@archivetvmus71 classic theme tune as well, sung by Dennis and co-written by his 'er indoors

[Antonio García Martínez (agm.eth) @antoniogm]

You can finally say this without being canceled: AI isn't creating a Cambrian explosion of apps, if anything it's holding app creation back.

Earlier tech waves had 'the mythical man-month'. Our generation has 'the mythical AI engineer' who magically turns enormous token usage into equally enormously-adopted products.

Well, where are the apps and new businesses then? Because compared to prior cycles (e.g. the mobile app boom starting in 2010 or so), right now seems positively sterile, app and UX-wise. Other than Claude or ChatGPT itself, name a new app you use now you weren't already using five years ago?

It's a truism of tech that throwing more people and time at a product often results in only lack of focus, confusion, and yet more code to support.

This is the parable of the company that over-raised and over-hired and grew too quickly, and now has lots of mediocre, weakly-adopted products, internal communication problems, distracted leadership, code bloat, technical debt...and so the spiral begins, which ends with an apologetic CEO post after some layoffs announcing "we're refocusing on our core customer".

Every tech company announcing they're either lowering token caps or shifting to lower-priced models is essentially saying: "we o̵v̵e̵r̵-̵h̵i̵r̵e̵d̵ over-spent on tokens, and are scaling back to focus on our core product" blah blah blah...same same.

It's the corporate version of someone using AI to write a long email, someone else using AI to summarize it, and both sides would have been better off just writing a shorter email. But now, even small companies can have that same problem thanks to AI.

I refuse to believe that an LLM prompt is the teleological endpoint of human interaction with computer intelligence. The fact we've apparently recrudesced to CLIs, like me farting around with RedHat 7.1 in 2001, feels like a step back. Another world here has to be possible, and while I have every faith (as someone as deep in AI psychosis as the next person) that AI can help get us out of it...just racking up tokens costs isn't how we get there.

The AI Jesus isn't coming to save us, human taste, discernment, and radical re-invention will. Like Kafka wrote in his notebooks: "The messiah will come only when he is no longer necessary; he will come only on the day after his arrival; he will come, not on the last day, but on the very last.”

[Muzoweri @muzoweri]

@snnneee 1955 - The Night of the Hunter: gothic nightmare
1927 - Sunrise: silent cinema dream
1946 - A Matter of Life and Death: cosmic wartime romance
1988 - Grave of the Fireflies: childhood under fire
1985 - Come and See: war as apocalypse

[Muzoweri @muzoweri]

@PathOfMen_ "Shyness is nice, and Shyness can stop you from doing all the things in life you'd like to"

[Paul Graham @paulg]

@mosasaurus27 YC looks for the same things in biotech startups as in other kinds of startups: founders with some kind of expertise and an idea that they, as domain experts, find intriguing, even if the rest of the world doesn't get it yet.

[Paul Graham @paulg]

How to Earn a Billion Dollars: https://t.co/WeWBUkKym6

[Muzoweri @muzoweri]

@aiwithmayank As the subtitle of the book says, this is for non-fiction. Don't follow this advice for fiction, and don't follow these bullet points without understanding the context.