Lucas
111 posts


Comparison demo without and with the Transitions skill
Demo page with more details, usage and commands
transitions.dev/skill
npx skills add jakubantalik/transitions.dev
English

Introducing: Browser Harness. A self-healing harness that can complete virtually any browser task. ♞
We got tired of browser frameworks restricting the LLM. So we removed the framework.
> Self-healing — edits helpers. py on the fly
> Direct CDP — one websocket to Chrome
> No framework, no rails, complete freedom
> Drop-in for Claude Code and Codex
I challenge anyone to find a task that DOESN'T work. I couldn't yet.🔥
100% open source ↓

English

@andrewchen 100% However, I would say it massivly boosts people good at context switching and multitasking.
English

@densumesh Very cool! I am curious if you benchmarked this in multi-lang like how well agents can find stuff if the content is in mixed languages.
I feel like there RAG, with vectors could still be a thing 🤔
English

to address the allegations, yes i forget to mention that OurModel is actually urModel, but we added some Optimizations
Matan Grinberg@matanSF
excited to annouce the latest scores of OurAgent on OurAgentBench: 1. OurAgent 2. YourAgent arxiv paper in bio
English

I give it about 3-6 months before any kind of skills.md file is also pointless.
The same thing happened to vector databases and langchain and every other 'product' built in the narrowing gap of model competencies.
@levelsio@levelsio
Thank god MCP is dead Just as useless of an idea as LLMs.txt was It's all dumb abstractions that AI doesn't need because AI's are as smart as humans so they can just use what was already there which is APIs
English

Migrate to GPT-5.4 with this Codex skill:
github.com/openai/skills/…
English

GPT-5.4 is here.
Native computer-use capabilities.
Up to 1M tokens of context in Codex and the API.
Best-in-class agentic coding for complex tasks.
Scalable tool search across larger ecosystems.
More efficient reasoning for long, tool-heavy workflows.
openai.com/index/introduc…
English

@petergostev Haha, I agree the urge to almost max out tokens, you feel it constantly
English

I have a new type of anxiety - what if I don't spend the tokens and they reset the limits??
Tibo@thsottiaux
This codex issue is now fully resolved and stable for the last couple of hours. You have come to expect it, but yes, that means we will be reseting rate limits in a bit. Enjoy.
English

@andrew_k_chan @ericzakariasson @aye_aye_kaplan @leerob Hey started happing again can u DM me @andrew_k_chan
English

@TheLicke @ericzakariasson @aye_aye_kaplan @leerob Hey! I work on the client at Cursor. If you DM more details, I can take a look
English

@hahnbeelee I am curious how grep works when the content is in multiple languages. Or the question is in one and the docs in another.
Have you done any tests regarding this? In my head semantic search should work better for that, no?
English

semantic search is dead. file-based context retrieval is far better.
let us know what you think and if there's more you want to see from our AI assistant
Mintlify@mintlify
The assistant just got a lot smarter It's now much faster and more precise What's new: → Grep retrieval (more precise than semantic search) → Multi-modal (ask questions about images) → OpenAPI search (ask about a specific API endpoint)
English

10 MCP servers worth installing right now:
1. filesystem — read/write local files
2. github — full repo operations
3. postgres — query your database
4. brave-search — live web search
5. puppeteer — browser automation
6. slack — send messages programmatically
7. notion — read/write your workspace
8. linear — manage issues + projects
9. memory — persistent key-value store
10. sequential-thinking — structured reasoning
All official. All free. All one command to install.
MCP is the USB-C of AI tools.
Everything plugs in the same way.
English










