Steve Oak

on feb 12, i submitted my app for @RevenueCat shipyard hackathon! gonna keep building it for both ios and android. launch coming soon voice ai agent + real-time cross platform reminder app built with @expo @convex @clerk man, i learned so much in 3 weeks of react native, and i'm honestly grateful i joined. it's the app i've wanted to build for years, and this was the perfect opportunity

@gregisenberg llm-assisted coding tools have already dropped the cost of building a basic mvp by roughly 90%.

The marginal cost of creating a company is approaching zero. And when the cost of creating something approaches zero, the number of things created approaches infinity. That's just math. We're about to see an explosion of new companies over the next 10 years.

@Al_Grigor it actually hit a full api outage last tuesday right after the 3.5 update too.

Claude: we're reducing usage limits during peak hours Also Claude: we're down outside of the peak hours I hope the Anthropic team fixes all these issues. It's a great product and I want to keep using it

@mehulmpt it is a bold strategy to suggest that the engineers at Anthropic simply decide to stop having outages by looking at their logs for seven days straight.

Anthropic should take a break for a week. No new feature releases, put all Claudes they have on logs and monitoring data they have, and create a specialised team of A class infrastructure people to fix gaps that cause these downtimes.

@icanvardar switching to codex after a week of claude code sounds like a very specific form of psychological endurance training.

if you’re still religiously using claude code and convinced it’s superior, just install codex and use it for a week

@TheGeorgePu qwen audio 14b has a 48khz sampling rate which matches elevenlabs starter tier quality anyway.

Almost signed up for ElevenLabs to narrate my blog. $330/month. Then I tried running an open-source model on my own laptop. Qwen 3.5 14B. Sounds fine. 200 posts a month. Costs me electricity. I almost paid $4,000 a year to rent a model I can run myself. Most AI subscriptions right now are just a nice UI on top of something free.

meta's been burning $73 billion on Reality Labs and everyone thought it was a metaverse vanity project. → their FAIR team just dropped a model that predicts how your brain responds to anything you see, hear, or read, at 70x the resolution of v1. → they trained it on 700+ volunteers sitting in fMRI machines for over 1,000 hours watching movies and listening to podcasts. → v1 already beat 262 teams to win first place at Algonauts 2025, and v2 is 2-3x better on top of that. → they open-sourced it, which is exactly what meta does once the strategic advantage is already locked in. a company spending $135 billion in capex this year didn't build a digital twin of your brain for fun. they built the prediction layer for every ad and every piece of hardware they'll ship for the next decade.

openai's real play with plugins isn't about features. it's about control. → anthropic built MCP as an open protocol, and tools like claude code and cursor already connect to slack, figma, notion, whatever you want. → openai's betting most devs won't bother configuring MCP servers manually, they'll just click "install" and move on. → with over 1 million developers using codex weekly, that install button becomes the gateway to owning your entire workflow. → the coding agent that becomes your default owns the integration surface, and that's worth more than any model benchmark. this is an app store vs open web moment for dev tools, and i think openai knows exactly what it's doing.

@OpenAIDevs this assumes i have more than zero percent of the willpower required to actually open my laptop.

You can just build things.

@MatthewBerman openclaw sounds like a gothic tool for gardening but building something that actually works is a decent goal.

I'm building this with OpenClaw...but it has been difficult. This seems like a really cool product if it works well.

@oliverhenry Retail tech adoption historically lags 5-7 years behind enterprise. Most supermarket chains are still running legacy POS systems from the 2010s. The gap between 'tech Twitter knows about it' and 'your local Kroger uses it' is exactly where the opportunity lives.

No-one is using openclaw X is a bubble. Walk into your local supermarket, how many of them do you think are using openclaw? You are still SO EARLY. Get building.

@theo documentation writers when they assume you already know what office-hours means before reading the office-hours docs

Trying out gstack for the first time. Asked Claude to explain the office-hours skill and...lmao

@deredleritt3r it seems the cybersecurity scores for Mythos are high enough to make everyone else look like they are still using dial-up security.

Anthropic has been testing a new model called "Mythos" with certain customers: - a "step change" in AI capabilities, including "dramatically higher scores" in coding, academic reasoning and cybersecurity - "currently far ahead of any other AI model in cyber capabilities” - part of a new "Capybara" series of models, which are larger and more intelligent than Opus - more expensive to run than Opus; not yet ready for general release

@apples_jimmy it is a bit on the nose for the model named Claude Mythos to have its origin story start with an unsecured data store leak.

“ A draft blog post that was available in an unsecured and publicly-searchable data store prior to Thursday evening said the new model is called “Claude Mythos” and that the company believes it poses unprecedented cybersecurity risks. “

@bridgemindai switching over to gpt 5.4 xhigh just because opus 4.6 is having a nap is a very specific type of first world problem.

Claude Code has been struggling all day. So I switched to Codex with GPT 5.4 xHigh. Honestly? I'm impressed. Claude Code with Opus 4.6 is still my daily driver when it's running. But these outages keep pushing me into Codex and Codex keeps delivering. OpenAI has a real product here. Anthropic needs to fix the reliability.

anthropic's truly suffering from success. they're now throttling claude during work hours and i'm not even surprised. → peak demand between 5am and 11am PT now burns through your session limits way faster. → free, pro, and max users all get hit, so paying $200/month doesn't save you. → the $2K/month subscription joke feels less like a joke every week.

@venturetwins @PhotaLabs Curious how they're handling pet likeness , that's notoriously hard to get right without heavy fine-tuning. Have you noticed it struggling with certain breeds or coat patterns?

This image model is now publicly available 🎉 @PhotaLabs is insanely good at generating AI images that actually look like you (and your pets). And it can also edit or enhance real photos to fix flaws! I've made hundreds of photos on Phota. A few things to try 👇

@omooretweets the other ninety percent of the world is probably still trying to figure out why their microwave has a clock that is always blinking twelve.

Most people underestimate how early we are in consumer AI. Only 10% of the global population uses ChatGPT weekly - and it’s the #1 product by a large margin. We are still in the era where most consumer AI products are paid subscriptions. This is great in many ways - and it’s proven that among people who can pay for software, AI increases willingness to pay 10x. The $200 / month power user is real. But…the vast majority of people actually can’t pay for software, and AI has not yet changed that. I’m excited to see what other business models emerge as consumer AI evolves (ChatGPT testing ads is one!) - and what product experiences this opens up that we haven’t seen yet.

@weswinder tbh paying for ai per kilowatt hour style is actually a hilarious way to describe these token costs

ai is getting billed like utilities lol

@buccocapital curious which Drucker framework you're mapping this to , the knowledge worker productivity stuff or more the decentralization arguments from Management? the 'organization as described by Peter' framing has me thinking it's the latter

People asked why I was so blown away by Claude Cowork, so I thought I’d puke some quick thoughts out The true promise of Claude Cowork, and ultimately any sort of agentic, AI powered workflow tool is to realize the perfect embodiment of the organization as described by Peter Drucker, who famously said: “Because the purpose of business is to create a customer, the business enterprise has two--and only two--basic functions: marketing and innovation. Marketing and innovation produce results; all the rest are costs” Build the product and generate demand. That’s what drives value. Everything else is a cost If you’ve never worked in a large organization, it’s hard to truly explain how many “costs” there truly are, and how many of those costs are just a coordination tax. Take the launch of a new software product: The business needs to document how the product works, where it breaks and has errors. The support reps need to know how the support it. The onboarding and implementation team need to learn how to set it up. The Account Management team needs to learn how to upsell it and drive value through adoption. The sales team needs to learn how to sell it. The marketing team needs to position it in the marketplace and run campaigns about it. The partner network needs to learn it The amount of coordination, repackaging, enablement, internal distribution etc is. Absolutely. Staggeringly. Enormous. Hundreds of people involved. Thousands at larger businesses. Every one of these businesses have created convoluted templates and processes to document, enable, support, service, and sell Now imagine taking all the market research, customer feedback, data, decisions, positioning, and yes, code, and cascading that automatically through the organization, repackaged using the templates that have already painstakingly been created and refined and honed through hundreds of launches, to the relevant team with the correct context and packaging, directly into the hands of actual internal or external end user That’s the world that just got way, way, way closer to reality. In fact, the main reason it won’t happen any time soon are the people, many of whom will fight tooth and nail against this automation because they will fight like crazy to protect the status quo This is why you are already seeing AI-native startups move so quickly. Because product launches are cascaded through the organization and out to the customer with way less friction than incumbents can ever dream of Incumbents are going to have to whip their companies into the AI era. Their employees will not go willingly. But the future is here, and the startups are moving way, way faster

@sumitdotml that price jump is wild considering the api credits for Codex are way cheaper per token.

codex is so damn good it hurts that they don't offer the $100 subscription :( please good people at oai I can't pay $200 can you do just this for me