Hayden™ retweeted
after switching our approach to building a "FirmOS", I'm increasingly confident that the future of agent collaboration within organizational settings is going to be a "company operating system" or CompanyOS, designed for agent and human collaboration
instead of a "company brain" that holds everything, the OS is for storing rules, processes, decisions made, which tools to use and how. it can mirror some data, but the source of truth is often best held in your existing tools (CRM, etc.). it's continuously updated and used by your agents
i think of this as a "flipped playbook" where instead of writing down the company rules first and following them (which is kind of like waterfall development), you're constantly updating the playbook as you decide or learn new things day-to-day running the business
Hayden™ retweeted
I built a dream logger.
For all those dreams I forget five minutes after waking up.
This little @M5Stack CoreS3-SE turns dreams into images using the @OpenAIDevs image-gen API. 🧵
I have been working for a while on a setup that takes tens of millions of rows of really wide tables (billions of cells) in the cloud and executes purposefully suboptimal (AI-generated) arbitrary sql queries on it. there are a lot of conditional and case-based steps where a couple of tables compute and then the output of that table becomes the input of the next case-based conditional. also the data changes every day, so every table is versioned and a query has to know exactly which version it ran on.
I was storing the input tables in parquet chunks on cloud buckets and running the compute on low concurrency auto scaling cloud run workers with 8 cpu cores, 32 gb of ram, and a bunch of optimizers and columnar compute. once i sniffed out and got rid of all of the dumb single-threaded python that gpt 5.6 had put in last month, we actually got really fast at the compute and were able to do most of the coherently written sql chains on billions of cells in a few seconds.
anyways, the jobs were still taking several minutes. but this weekend I totally changed the foundation of how this is all built and it got like 20-100x faster end to end.
if you can guess what was making it slow, or what the new setup is, or help me make it 100x faster from here, or generally just want to geek out about these things, I can offer Pokémon Center: 30th Celebration ETB's for considering a new full-time job opportunity in NYC.
2
5
144
Celebrate Columbus however you want, so long as you don't think for a second about trading a sovereign or corporate credit instrument today.
Today's not the day for that, pal. That's not what He would have wanted.
105
🤖 Made with AI
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…
1
226
wow didn't actually expect this to be true
4
221
5.6 Sol did 1B+ tokens for me in 20 hours since release.
Total workhorse, highly productive model, not seeing a ton of token waste. Next generation vibes are strong.
1
5
160
Hayden™ retweeted
I cannot emphasize enough how good it feels to have access to a next generation model again.
thank god. azure foundry is a nightmare.
we have updated our partnership with microsoft.
microsoft will remain our primary cloud partner, but we are now able to make our products and services available across all clouds.
will continue to provide them with models and products until 2032, and a revenue share through 2030.
175
the millisecond someone builds this everyone will expect it to be free.
often when you hear "I would pay a lot of money for" it sounds like a representation of value but really it's a representation of desire.
they are not the same.
if you disagree, build this product and let me know how many people are willing to actually "pay a lot of money" for it.
I would pay a lot of money for some AI solution that clips the 10 most interesting minutes from the podcasts I follow and gives me an hour or two of the best content each week
1
2
203
there is a lot of hate on Higs but tbf they have an amazing product offering and the haters are in the wrong
This quoted post is unavailable.
1
140
Hayden™ retweeted
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.





















