Facundo García Martoni

@rezoundous That’s how all hiring is supposed to work Tyler

@thsottiaux Yes! Babysitting a PR or running automations are perfect use cases for this. I don't need fast or even normal speed for these tasks in day-to-day work.

Should we bring batch compute to codex? Aka /slow mode

Bob McGrew has a framework I keep thinking about: in the AI future there are only two jobs. The Lone Genius and the Manager. That's it. Everything else gets absorbed. The Lone Genius is the person sitting alone at a computer, amplified 1000x by AI. One person with taste, vision, and relentless focus who can now do what used to take a team of 50. The Manager is the person who becomes CEO of their own "firm" where most of the employees are AI agents. They define the goals. They decide what matters. They coordinate. The AI does the execution. The Marxists will hear "two jobs" and panic. "What about everyone else?!" But here's what they're missing: AI doesn't shrink these two categories. It explodes them open. More people get to be geniuses. More people get to be managers. The barrier to entry for both just collapsed. What actually gets eliminated? David Graeber called them "bullshit jobs." Graeber was no libertarian! He inspired Occupy Wall Street. His words: "Huge swaths of people spend their entire working lives performing tasks they secretly believe don't really need to be performed. The moral and spiritual damage that comes from this situation is profound. It is a scar across our collective soul." Graeber said bullshit jobs are "a form of spiritual violence directed at the essence of what it means to be a human being." They induce "hopelessness, depression, and self-loathing." This is who the left should be fighting for. Not to preserve those jobs. To liberate people from them and give them better ones. The dirty secret of the modern economy: millions of people sit in roles so pointless that even they can't justify their existence. Compliance layers. Reporting layers. Coordination layers. Meeting-about-the-meeting layers. They know it's meaningless. It eats them alive. AI eats those layers. Good. That's a jailbreak. What I love about Bob's framework is where it points. The Lone Genius used to require a PhD, a lab, institutional backing. Now a 19-year-old with taste and Codex can ship what took a research team a year. The genius bottleneck was never talent. It was access. The Manager used to mean you needed to hire 50 people, raise money, build an org chart. Now you can orchestrate a fleet of AI agents from your laptop. The management bottleneck was never skill. It was capital. AI doesn't concentrate genius and management into fewer hands. It distributes them into more hands. The working class kid in West Virginia. The single mom in Ohio. The 55-year-old who got laid off and now builds software for the first time. Those are some of Bob's future geniuses and managers. The best founders I see at YC are already living this. They toggle between both modes in the same day. Morning: lone genius, creative insight, the thing nobody else sees. Afternoon: manager, spinning up agents, steering, shipping. The cycle time between genius and manager IS the new productivity metric. So when someone tells you AI means "only two jobs and everyone else starves," quote Graeber to them, they’ll get it. Graeber knew the real violence was making people do meaningless work and pretending it was dignity. AI ends that. More genius. More agency. Fewer spiritual prisons.

Awesome share Jason. Sharing some thoughts that came to me.. Everyone indeed is saying productivity is the "worst" category. But that's extremely short sighted. Over time, there won't be winning and losing categories. In fact, there won't be categories at all. All software will converge. This is the belief and strategy we started ClickUp and were laughed out of the room too many times to count. But now, everything IS possible. If you zoom far enough in the future, everything is the only thing that will be around as far as large companies go. Every company can now build any product they want in nearly zero time relative to the past. So what will happen? Naturally, every company will build MORE software not less. Over time, this leads to all software converging. In the future you will buy all of your software (and AI) from one vendor. ClickUp will be the first. The primary 'why' behind our 100x ORG is actually way more simple... We are a PRODUCTIVITY company. Our MISSION is to make the world more PRODUCTIVE. I've had FIVE near death experiences. Each one makes me more obsessed with productivity. We provide productivity to our customers - so we MUST be at the forefront of productivity ourselves. AI exists for one reason only... PRODUCTIVITY. You use AI to be more productive. Everyone does. I very clearly see the AI future that will happen whether we do this first or we do this last. It'll happen either way. There is an organizational leapfrog available. We will look back a year from now and this won't be so controversial anymore. While we're at it, here's more futuristic thoughts that came to me when writing this.... PREDICTIONS 1) Software categories with historically little to no competition will get crushed - as AI coding allows new competitors into spaces that previously had none. 2) Yes, new competition will arrive in horizontal software as well. But it doesn't mater. The horizontal winners today are already used to thousands of competitors, thousands more won't change anything materially. Horizontal platforms that can execute quick enough win big in the short-term. 3) AI is stuck in single player mode today. That's already starting to change. Humans are now engaging with their agents and vertical platforms inside of your chat and work platforms. This is growing exponentially as we see it. Horizontal platforms unlock multiplayer AI - where humans and agents work together seamlessly. The same reason these categories exist in the first place, to enable human collaboration - are the reason they'll thrive with human engagement next (except for tech market, more on that below). 4) Vertical software dies. Horizontal software that's built in a compounding way (we call it 'Converged Software') will replace all vertical software using horizontal platform (personalized by AI). 5) Vibe coding is highly inefficient. I'm not even talking about your time/cost - I simply mean from first principles. Building apps 0 to 1 is insane. Vibe coding should be starting with .9 then going to 1. That last mile is all that's needed. This will change soon. 6) Additionally, building your own software is a builder thing. It's not a normal person thing. The vast majority of potential customers world-wide will never build their own software even if it was one click to do so. And they shouldn't, that's not what they do. 7) Chasing tech as your market is a race to the bottom. Yet most SF investors and companies seem to be chasing the exact same market. Yes, the $$ are there today - they are early adopters by definition. The rest of the world outside of X has never used an Agent, largely speaking. They've never used Anthropic, even if they've heard the name. Our market is 85% non-tech. This is where the real opportunity for delivering AI value is. More to come soon. 8) Everyone talks about a 'moat'. And the moat definition and prediction keeps changing doesn't it? I don't believe any of the moats that are accepted as moats today. Instead, the only moats in software of the future (outside of eng/tech market) are simply: (a) CONTEXT (in real time, injected in harness) AND (b) HUMAN ENGAGEMENT. Everything is changing faster than we can check twitter. I may be wrong about these predictions but one thing is for sure - everything we believe to be true today will not be true in the future.

@aidenybai I was trying to play with react.review yesterday, but it's still stuck here. Anything wrong? 🤔

@jskoiz @OpenAIDevs @thsottiaux this is important 👀

Context Indicator is missing in current Codex App you're not going crazy FYI. I see quite a few new issues opened in github so Im sure team is aware.

A mistake that cost me 5 years: Thinking preparation was progress. Reading every book. Taking every course. Planning every detail. Meanwhile, someone dumber than me started badly and figured it out. Preparation feels productive but it's often just fear dressed up as strategy. You learn to swim by getting in the water, not by studying water.

Excited to speak about AI in the Puna Tech's main track in Salta on May 30th!! Thanks for the invite @facundopadilla_ 💪 I'll give a talk on how to use AI to be as productive as possible for software engineering. punatech.ar/cronograma

@WaynaWaver @DamianCatanzaro No debería en cuanto al funcionamiento core, pero la app siempre va a estar un paso adelante en mac porque los developers de Codex (y de todo SF) codean en mac

@facmartoni @DamianCatanzaro Dices que hay diferencia entre usar códex en windows que en Mac?

Meti 2 prompts con GPT 5.5 en OpenCode y se me fumo 20% del daily, en esta casa se sigue usando 5.4 para desarrollar, bueno bonito y barato

OpenCode: the cool project Cursor: the underdog Claude: the expensive option Codex: the 🐐.

@DamianCatanzaro El CLI está muy bien pero la app nativa sobre todo en macOS le saca todo el jugo. Podés usarlo en xhigh en seguidilla todo el día y no vas a cansar el plan de $100

@facmartoni Decís con el CLI de Codex? Tendré que probarlo, OpenCode hace muy buen laburo

Unpopular opinion: I don’t care if most web apps look the same. All I care about is whether it does what it says and does it fast. Make it fast. Make the UX obvious. Put the right things in the right place and little to no animations.

The new Cofounder site is literally a step-by-step guide on how to start a company. When i started my first company there was all sorts of stuff i had to learn and all of the guides were SEO maxxed slop. Now there's a real one written by experts and u can download it.

@nicoproducto O las dos al mismo tiempo

Chicos no pueden ser todos devs, alguno tiene que estar pensando estratégicamente en el producto

@tomasmalamud @lapyme_ar Aguante Shadcn Tomi! Nuestros dashboards parece idénticos jajajaja es lo mejor que nos pasó

Lanzando dashboard personalizable en @lapyme_ar! Ahora cualquiera puede armar su propio dashboard con los reportes creados en el armador de reportes personalizados (donde pueden definir la visualización, métricas, dimensiones y filtros). Sentando las bases para una feature tipo Pulse para el futuro agente de La Pyme.

This works really well btw, at the end of your query ask your LLM to "structure your response as HTML", then view the generated file in your browser. I've also had some success asking the LLM to present its output as slideshows, etc. More generally, imo audio is the human-preferred input to AIs but vision (images/animations/video) is the preferred output from them. Around a ~third of our brains are a massively parallel processor dedicated to vision, it is the 10-lane superhighway of information into brain. As AI improves, I think we'll see a progression that takes advantage: 1) raw text (hard/effortful to read) 2) markdown (bold, italic, headings, tables, a bit easier on the eyes) <-- current default 3) HTML (still procedural with underlying code, but a lot more flexibility on the graphics, layout, even interactivity) <-- early but forming new good default ...4,5,6,... n) interactive neural videos/simulations Imo the extrapolation (though the technology doesn't exist just yet) ends in some kind of interactive videos generated directly by a diffusion neural net. Many open questions as to how exact/procedural "Software 1.0" artifacts (e.g. interactive simulations) may be woven together with neural artifacts (diffusion grids), but generally something in the direction of the recently viral x.com/zan2434/status… There are also improvements necessary and pending at the input. Audio nor text nor video alone are not enough, e.g. I feel a need to point/gesture to things on the screen, similar to all the things you would do with a person physically next to you and your computer screen. TLDR The input/output mind meld between humans and AIs is ongoing and there is a lot of work to do and significant progress to be made, way before jumping all the way into neuralink-esque BCIs and all that. For what's worth exploring at the current stage, hot tip try ask for HTML.

@thdxr So needed! In my startup we only deal with one of the biggest providers, and omg that API is so freaking bad

we're working on a library to abstract over all the llm providers there's very few teams that have dealt with the quirks between providers at the scale we have it's written in effect but will also have a vanilla api progress is in the opencode repo under packages/llm