Come hang out in the land of innovators, with the best showcasing the future.

Why even non-coding agents need bash I've done dozens of calls with companies making general agents over the past few weeks and my advice generally boils down to: "use the bash tool more" Here's a concrete example from my email agent:

Design and product UX is such underrated alpha

Sam altman wanting to train his next model

AI Video Creator / Editor === Next Software Engineer?

the only way I would love to be remembered as.

It feels like there is too much pressure on big AI companies that they have forgotten how to experiment

Just read the recent a16z newsletter and damnnnn, these guys cooked! > Generalists which can understand all the spheres will get more important, as they decide the last 1% > Legal, Video Editors (and certain extent software as well) will get more valuable as people end up

In the Smol Training Playbook, I tried to survey the state of popular post-training frameworks. Let me know if I missed any and I'll add them to the list!

We want to sell muffins to AI. That's when it will be an AI bubble.

Everyone wants to “get into ML” until they meet: • Data pipelines • Distributed training • Infra costs • Latency nightmares If you’re entering tech, learn distributed systems first. ML becomes way less scary after.

Dear “15-18 yo founder”s sending me DMs, don’t. Go and hug your parents, fall in love, eat chocolate cereal for breakfast, read poetry. Nobody will give you back these years. And sure, do your homework and learn math and code if that feels fun. But stop building SaaS and

Work hard, if you really have to. But know why you’re working hard. And keep re-evaluating those reasons from time to time. Fetishization of hard work creates toxic culture, and robs humanity of everything that’s important for a happy life - health, play and relationships.

I have been testing out MCP Servers and CLI tools. I think CLI beats MCP hands down. MCP restricts one to specific tool calls and increases token consumption by huge numbers. CLI Tools >>>>>> Skills >> MCP Servers

If you are a recruiter, do not discourage use of LLM, instead ask them to make it one shot LLM execution. Recently while working on a backend task, I created a long spec of backend architecture and then paired it with another agent whose role was to test the code written and

When your AI agent gets a major power-up 🔋 See how the Chrome DevTools MCP server works → goo.gle/3KU63q2

People don’t like AI generated content / code unless a person has put some love and effort into it. The person is still very important, perhaps more important than before