Keep on to blur preview images; turn off to show them clearly
Every Saturday, I send 175,000+ entrepreneurs the strategies I use to turn expertise into income. Read it, apply it, and your business gets better. Always 4 minutes (or less) to read. Free to join here: https://t.co/idNByMf6vn
The $10M Solopreneur | Helping 100,000+ experts turn their expertise into income.
Sergey was definitely feeling the challenge from OpenAI. You can almost hear him saying, “We invented the Transformer, you took Ilya from us, game on.” That challenge reignited his passion to return and win. Competition is a powerful force for technological advancement.
Co-founder & CTO @hyperbolic_labs cooking fun AI systems. Prev: OctoAI (acquired by @nvidia) building Apache TVM, PhD @ University of Washington.
RT @alexolegimas: Thank you Ryan. This is what non-Jews probably don’t realize: to practice Judaism essentially anywhere on the globe now r…
Market Design/Entrepreneurship Professor @HarvardHBS & Faculty Affiliate @Harvard Economics; Research @a16zcrypto; Editor @restatjournal; Econ @Quora; … | #QED
I’m getting more and more convinced that the Unix tool approach of having a bunch of focused, composable functional units that can be used in isolation or as part of a larger pipeline is also the best approach for tooling for coding agents. The problem with trying to make a big unified system that does everything is that people have their own workflows and ways of doing development, and it’s usually too hard to try to make a one-size-fits-all project that can accommodate that without it turning into sprawling complexity that doesn’t work well in practice. So I have one tool for agent mail, one tool for task management (Steve Yegge’s beads), one tool for task selection (bv), one tool for past history search (cass), one tool for polyglot linting and bug-catching (ubs), one for sensitive command handling (slb), one for managing tmux and agent sessions (ntm), one for memory (csm), etc. And you can use one of them, some of them, or all of them. And they’re partially integrated with each other, but always optionally. So slb can use agent mail if you have it set up, but it also works standalone. And ntm can display info from bv but it isn’t forced. They become like little Lego blocks that you can use to build up whatever system you want, and it becomes easy to create your own little tools to add the functionality you want. And then your AGENTS dot md file sort of becomes like an operating system where you “install” the tools in the agent’s working memory and configure them by describing how, when, and why to use them (weirdly, you now have to worry about persuading the machine to use the tools!) So you can have all your tools set up on your machine but only activate specific ones depending on the project by including just the relevant blurbs explaining the tools you want to use.
Former Quant Investor, now building @lumera (formerly called Pastel Network) | My Open Source Projects: https://t.co/9qbOCDlaqM
Fresh off the federal building. The Bentley coupe with the convertible ceiling. Im the black widower, call me mr white. I don been through it all, shoot outs and fist fights.
Founder | Author | Speaker Building @beltstripe. Healtech/EdTech/Agric I'm Not The Man Of Your Dreams. Your Imagination Wasn't This Great.
@krea_ai If you want to turn your images into video, I would recommend Kling 2.5. You can keep prompts relatively simple. I did things like "woman taking video on movie set, turns to talk to characters behind her). (I also did the video generation on @krea_ai)
Partner @a16z AI 🤖 and twin to @omooretweets | Investor in @elevenlabsio, @krea_ai, @bfl_ml, @hedra_labs, @wabi, @WaveFormsAI, @ViggleAI, & more
