Kennedy

Plan mode feels unnecessary with Codex - it is that good. But I still plan large features.

It is starting to feel agentic with about 9+ tools. No UI yet, so I made a simple trace viewer to help understand what is happening. Next, I'll be adding session management.

Many understand a dialect but may lack the confidence to communicate in it. The same can be said about programming languages in this agentic era, where one may learn solely to review AI output. For example, I learn Rust to keep up with AI. I’m a read-only Rustacean.

@_trish_xD go back to jQuery? No thanks!

React is proof that developers would rather deal with complexity than learn vanilla JavaScript.

It made me think about a future where apps provide tools, assets, endpoints, and documentation, while relying on browsers or agents to create dynamic, accessible interfaces tailored for users. We could create a highly composable (permissionless?) internet.

I don't know what to look for in TUIs, but I need one for Hand. Bubble Tea seems like a good one, especially with the MVU pattern + lipgloss.

Sam Altman gave up direct control of OpenAI's safety and security teams, moving safety under CRO Mark Chen and security under president Greg Brockman, so he can focus on raising money, supply chains and building data centers at a massive scale OpenAI finished pretraining its next big model called "Spud" and expects a very strong model in a few weeks that can accelerate the economy, and is shutting down the Sora video app and API to free up computing power for it and shelved plans to bring video features into ChatGPT Sora research will shift to long-term world simulation focused on robotics, OpenAI renamed its product org to "AGI Deployment", and Sam noted things are moving faster than many expected

Your body has an enzyme that can stop your cells from aging. You produce it as an embryo, and then your body shuts it off. On purpose. Because if your cells kept using it, they’d never stop dividing, and that’s literally what cancer is. About 85% of human cancers work by switching this enzyme back on. Lobsters never switch it off. Ever. They produce this thing, telomerase, in every single organ their entire lives. A 1998 lab study found it running in their liver, their heart, their muscles, everywhere. What it does is simple: every time a cell copies itself, the tips of your DNA fray a little, like the ends of a shoelace losing its plastic cap. Telomerase rebuilds those caps. In your body, most cells can’t do that, so they wear down and eventually die. That’s aging. Lobster cells just… keep going. And here’s the part that messes with me. They actually get more fertile as they age. A 1-pound female carries about 8,000 eggs. A 9-pound female can carry over 100,000. The grandmother in this video is hauling 70,000–80,000 eggs. In almost every other animal, reproduction slows down and stops as the body breaks down. Lobsters flip that completely. The older they get, the more babies they can have. They do still die. Just not from aging. Lobsters grow by cracking out of their entire shell and building a new one from scratch. Imagine doing that every year or two, and each time it costs more energy because you’re bigger. Eventually the molt itself is what kills them. Exhaustion. Or infection while the new shell is still soft. The oldest one anyone’s found, a 20-pound male named George, was about 140 years old when he got pulled up off Newfoundland in 2008. PETA got involved, and he was dropped back into the water near Kennebunkport, Maine. This female getting released isn’t random feel-good content, either. Maine has made it illegal to keep egg-bearing females since 1872. When a fisherman catches one, he cuts a V-shaped notch into her tail flipper before tossing her back. That notch makes her permanently off-limits, even years later when she’s caught again without visible eggs. A 2018 study found that V-notched females produce 9 times more eggs over their lifetime than unmarked ones. And the survival odds for lobster babies are terrible: out of every 50,000 eggs, maybe 2 make it to a size you could legally catch. One big, old, egg-loaded female like this one is worth more to the species than thousands of younger lobsters combined.

Owning the SOTA model and infrastructure doesn't mean they will build the killer application.

RT @soraofficialapp: We’re saying goodbye to Sora. To everyone who created with Sora, shared it, and built community around it: thank you.…

I mean, this is already happening. The fact that we are on X sharing ideas, not on our personal websites, proves it. The open internet will be lost to agents.

@ileke_airende Yes! lol. I’m enjoying building this and will be sharing more, including a GitHub link.

@ncodes Cooking again🔥 We wait!

Currently working on the foundation for my personal agent. I'm calling it Hand. Personal AI assistants remind me of Daemons in His Dark Materials - They follow you everywhere, are intelligent and know all about you. Small steps.

Introducing the MiniMax Token Plan: First All-Modality API Subscription Flat-rate API access to MiniMax's leading text, speech, music, video, and image models. Stop juggling multiple unpredictable bills for different modalities. One key. One predictable bill. All modalities.

I have come to the same conclusion. You can outsource your thinking but you cannot outsource your understanding

Code is an output. Nature is healing. For too long we treated code as input. We glorified it, hand-formatted it, prettified it, obsessed over it. We built sophisticated GUIs to write it in: IDEs. We syntax-highlit, tree-sat, mini-mapped the code. Keyboard triggers, inline autocompletes, ghost text. “What color scheme is that?” We stayed up debating the ideal length of APIs and function bodies. Is this API going to look nice enough for another human to read? We’re now turning our attention to the true inputs. Requirements, specs, feedback, design inspiration. Crucially: production inputs. Our coding agents need to understand how your users are experiencing your application, what errors they’re running into, and turn *that* into code. We will inevitably glorify code less, as well as coders. The best engineers I’ve worked with always saw code as a means to an end anyway. An output that’s bound to soon be transformed again.

Today’s objective is to build a working loop with environment and context management. Most interesting things run in a loop.

I’ll be spending the next few weeks researching and developing a personal agent in Go. I expect to enjoy the process.