Marco Kobayashi

[Marco Kobayashi @marcodlk]

@bentlegen @pidotdev Do you see this enabling text search as an extension?

[Thariq @trq212]

been asking others at Anthropic how they stay in the loop with Claude and fully understand the work being done

this is one of my favorites from Suzanne:

[Nathan Baschez @nbaschez]

Introducing Roughdraft!

A new open source project designed to make collaboration with agents better.

The idea is to bring commenting and suggested changes to markdown (e.g. plan docs) in a nice interface.

Free, local, etc.

👉 https://t.co/J3YOOpL5ES 👈

[Yiwei Ho @1weiho]

Introducing open-slide - The slide framework built for agents.

Prompt your agent, get a polished deck.

$ npx @⁠open-slide/cli init

👇

[Google DeepMind @GoogleDeepMind]

Meet Gemma 4: our new family of open models you can run on your own hardware.

Built for advanced reasoning and agentic workflows, we’re releasing them under an Apache 2.0 license. Here’s what’s new 🧵

[Claude @claudeai]

You can now enable Claude to use your computer to complete tasks.

It opens your apps, navigates your browser, fills in spreadsheets—anything you'd do sitting at your desk.

Research preview in Claude Cowork and Claude Code, macOS only.

[Akshay 🚀 @akshay_pachaar]

https://t.co/SSSIK3BX4z

[Neethan Wu @neethanwu]

https://t.co/z7p1rL84KV

[sarah guo @saranormous]

Caught up with @karpathy for a new @NoPriorsPod: on the phase shift in engineering, AI psychosis, claws, AutoResearch, the opportunity for a SETI-at-Home like movement in AI, the model landscape, and second order effects

02:55 - What Capability Limits Remain?
06:15 - What Mastery of Coding Agents Looks Like
11:16 - Second Order Effects of Coding Agents
15:51 - Why AutoResearch
22:45 - Relevant Skills in the AI Era
28:25 - Model Speciation
32:30 - Collaboration Surfaces for Humans and AI
37:28 - Analysis of Jobs Market Data
48:25 - Open vs. Closed Source Models
53:51 - Autonomous Robotics and Atoms
1:00:59 - MicroGPT and Agentic Education
1:05:40 - End Thoughts

[lucas gelfond @gucaslelfond]

snowstorm hack, zerobrew is a drop-in brew replacement. borrowing principles from uv (concurrent downloads, content-addressable store), it’s ~5x faster cold and ~20x faster than homebrew.

try it out! https://t.co/TGzrq28zzQ

[Palantir @PalantirTech]

https://t.co/fLf0SLzHd5

[Excalidraw @excalidraw]

We've made text-to-diagram better.

Chat interface. Streaming. Smarter. Faster. Stronger.

[ℏεsam @Hesamation]

this 3b1b visualization of how LLMs work layer by layer is the true brain rot we never knew we needed.

[Stanislav Kozlovski @kozlovski]

Ever wanted to quickly diagnose a Kafka cluster without spending days combing through a ton of data?

The MIT Licensed kafka-analyzer connects to Kafka and spits out a concise HTML report of things that are good and things that are bad.

It checks things like:
⚖️ are your partition replicas evenly distributed?
🗜️ is compression enabled? (can save bandwidth costs 4-10x)
💀 do you have any dead consumer groups? (with inactive members)
🌏 are racks configured on your cluster? (i.e is it multi-AZ)
📊 is your cluster exporting JMX metrics?
🔢 does your replication factor make sense? (e.g not more than the broker count)
🚨 are there any under-replicated partitions?
📏 are your partition counts uniform, or do they vary wildly?
🔐 is authentication configured?
♾️ do any topics have infinite retention?

It works with every popular Kafka distribution - MSK, Confluent Cloud, Aiven, Redpanda, and Kafka.

It’s an awesome CLI that's super easy to use.

🐣 It’s still early days - I can think of a few things to polish - but the idea is really strong.

This is what the Kafka ecosystem has been missing—a dead-simple open source diagnostic tool! 💪

[Simon Willison @simonw]

I figured out how to add the official Playwright browser automation MCP to Claude Code. Run this before you start "claude":

claude mcp add playwright npx '@playwright/mcp@latest'

Now Claude Code can use a Chrome browser directly!

Here's my TIL: https://t.co/O5KcDe6U3M

[Tom Dörr @tom_doerr]

library for building AI agents with memory, knowledge, tools, reasoning

[Jessie Frazelle @jessfraz]

Under every single CAD application that businesses use is a thirty year old CAD kernel.

Even OnShape and Shapr3d, “modern” tools use Parasolid a 30 year old beast.

We, on the other hand, are building a new CAD kernel. To use GPUs not just for rendering but for the math.

Yes, open source and hobbyists have created some newer projects in this area but they aren’t used by businesses for production products. They lack features, UX, adoption, etc.

[Sewoong Oh @sewoong79]

We are releasing OpenDeepSearch (ODS), an open-source search agent that works with any LLM. When paired with DeepSeek-R1, ODS outperforms OpenAI’s specialized model for web search, GPT-4o-Search, on the challenging, multi-hop FRAMES benchmark from DeepMind (+9.7% accuracy).

[Pydantic @pydantic]

Pydantic Evals is a powerful evaluation framework designed to help you systematically test and evaluate the performance and accuracy of the systems you build, especially when working with LLMs.

[myles @themylesfiles]

holy damn, this is awesome. the new `--watch` command on @marimo_io paired with markdown extensions 🤯

```
marimo edit notebook\.md --watch
```