Ryan Schmidt

Pressure feels like a threat, but it's not. You feel pressure when your decisions matter, and people depend on you. It can feel uncomfortable at times, but it's also a privilege. When no one relies on you — when no one expects something from you — you're irrelevant. Pressure is a privilege. @shaneparrish

@BrendanEich is there a trick to re-enabling my brave wallet? Something changed over the last 6 months (or longer?) and now I’m forced to pay gas fees to set up a new wallet so I collect some funny money for the ads. I like Brave otherwise but it seems like I’m missing something.

@TheNathanRuff Talk to me about the bird on your arm

Departing Nashville… on my way to Georgia

@agazdecki Haha so true

Startup founders when they realize building the product was actually the easiest part:

@SawyerMerritt Link?

NEWS: Tesla has announced a Signature Edition Model S and Model X as a final goodbye for these two vehicles. • Price: $159,420 • Only 250 Model S & 100 Model X Signature Editions will be built. All Plaid variants. • Garnet Red exterior paint • Matching Garnet Red Door Handles • Carbon Ceramic Brakes with Gold Calipers • Gold Plaid Seat Badging + Gold Piping + Alcantara Seats • Gold Tesla T Badge • Signature Marked Door Sill • Signature Dash Badging with Unique Edition Numbering • Signature Liftgate Applique  • Rear Gold Plaid Badge • Interior Lighting Sequence • Signature Edition Key Fob • And more

@Hesamation Operating like a business.

AMD Senior AI Director confirms Claude has been nerfed. She analyzed Claude's session logs from Janurary to March: > median thinking dropped from ~2,200 to ~600 chars > API requests went up 80x from Feb to Mar. less thinking and failed attempts meaning more retries, burning more tokens, and spending more on tokens > reads-per-edit dropped from 6.6x → 2.0x. model stops researching code before touching it. > model tried to bail out or ask "should i continue" 173 times in 17 days (0 times before March 8). > self-contradiction in reasoning ("oh wait, actually...") tripled. > conventions like CLAUDE.md get ignored because there's less thinking budget to cross-check edits > 5pm and 7pm PST are the worst hours, late night is significantly better. this means the thinking allocation is most likely GPU-load-sensitive.

@pmddomingos Seems true

TL;DR: Top hacker calls Anthropic’s bluff.

@forgebitz This is why most “dev shops” will fail.

having mid developers is now going to slow your company down before a mid developer would do mid tasks like crud monkey work, forms, etc, which is fine, that needed to be done now you have a mid developer being able to generate massive amounts of slop code, but they don't know it's slop, so the better/senior devs need to review the slop creating a bottleneck of vibe coded slop

Sometimes people are so worried they’ll sound mean that you don’t understand what they mean.

@andrewchen Agreed. Vibe coding is the spreadsheet of 2026. Solves a problem, can work well, and can get you pretty far.

prediction: we'll soon view coding the way we view using spreadsheets today - a commonplace skill that every white collar worker is expected to have. Knowing how to code will sit alongside email, making slides, word processing, etc etc. It'll be <18 months before this is widespread in every job description Customer-facing employees will code as well as sell/market/support, so that they convert their domain expertise into repeatable workflows and software. We'll have an explosion of internal bespoke apps. But to complement all of this coding happening at the edges, we'll also have centrally expert teams of agentic coders who build infrastructure, make it secure/scalable, and create canonical software. These central teams will help scale agentic engineering. There's a spreadsheet metaphor here too -- yes, if you are an expert at spreadsheets, you write macros, build huge models, etc., we'll put you in a group of your own. It's called Finance. :) We'll have the same central teams to help manage the widespread use of coding tools throughout the org. You might ask, won't this be a mess? What happens in a world where everyone has many many variations of bespoke software? (It's already happening) Maybe! But I think it'll be fine, in the same way that it's fine to make a copy of a spreadsheet or deck. Making it easy to fork makes it easy to participate. But you might want an "official" forecast maintained by an official finance person, in the same way that there will be variations of canonical and bespoke software excited for the Gdrive of many many forks of disposable apps made and shared by my co-workers!!! 😂

@karpathy Yes. It was as if OpenClaw invented AI for those that hadn’t really used AI before. Those that had never seen AI do anything meaningful suddenly had their aha moment.

Someone recently suggested to me that the reason OpenClaw moment was so big is because it's the first time a large group of non-technical people (who otherwise only knew AI as synonymous with ChatGPT as a website) experienced the latest agentic models.

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.

@CliftonSellers Turning tribal knowledge into systems.

You got 5 words Sell me your service

@levie @a16z @stevesi It was like he didn't even hear it 😆

@ryanschmidt @a16z @stevesi He has years of practice with me 😂

Steven Sinofsky on why it's hard for AI to diffuse through firms: "Algorithmic thinking is really, really, really hard for the vast majority of people who have jobs… If you were to go into any person and ask them to create a flow chart for a particular thing that they have to go do, they would probably fail at producing that flow chart." "So within any organization, say doing a marketing plan… one person probably understands and could document the flow chart. So if you put one of these agents or this coworking tool in front of people… their ability to explain to it what to do is really, really limited." "You're basically just developing the next abstraction layer for how people interact… at each level of the abstraction layer, [it's] been a highly skilled, very specific individual within an organization… and then the little parts they build become little toollets… and some people can stitch together and some can't." @stevesi

@staysaasy Oh man I feel this so hard.

Friends outside of tech: lol copilot is dumb Friends in tech: I just bought iodine tablets and have made an offer on land upstate. My supplies of antibiotics and potable water are sufficient but I need to set up hydroponics to make it through the first few years.

@jjen_abel Incredibly accurate

startups, here's enterprise decoded ... 1. 'circle back' = dead 2. 'get it done asap' = 90 days 3. 'send me what I can share internally' = 3 bullets 4. 'this is exciting' = let me play a role where I can put show how I helped shape this for us 5. 'send me your paper' = i will send you mine back OR be prepared for 4-6 iterations on your paper 6. 'how much time is needed' = 30min OR less 7. 'I will bring my boss, who's joining on your side' = I expect you to match seniority -- ie. your founder 8. 'No questions from my side" = I have no idea what just happened 9. 'This price seems fair' = this falls far below what I had budgeted 10. 'Give me time' = stop following up mf-er, I'm on it

@TheNathanRuff 4. Joined Hampton 5. Met @ryanschmidt

@ryanschmidt We met in Hampton

I joined Trends Facebook group by The Hustle on Black Friday for $100 in 2020 That decision has the highest ROI of anything I’ve ever done 1. Made 1M+ off of contracts from people I met in the group. 3M+ when you include second degree network. 2. Met the guy who introduced me to @josiah_daves who’s Google ads changed my entire business and life 3. Became aware of @thesamparr

I don't know why you wouldn't be subscribed to @shaneparrish's newsletter. Always full of bangers like this one: Credibility is expensive because the bills never stop. You pay it in conversations no one overhears, in deals where you leave money on the table, in credit you give away. You pay for it every time you say the hard thing instead of the easy one. The strange part is that you're paying for years before anyone notices, and you can lose it all in an afternoon.

@chamath Setup one folder and tell the agents to store context files, not "memories". Everything gets organized nicely so that you can read it (I like Obsidian) and so that various agents can too. I use Codex, Claude, and OpenClaw with the same "vault". It's magical.

This may be a dumb question but I’ll ask it here anyways: I can’t find a good way for my various AI chats to automatically sync its conversation history into a structured knowledge base. So that as I update various chats from time to time and refine context, my knowledge base automatically grows with this new info.