@0xSero Have you solved the problem of sending screen shots to the agents? 🤔
I think currently this is the main bottleneck of having agents on a different server that you access with Tailscale
I have to say I agree with that.
Svelte is way better then react. I’ve been using react since 2017 and knows it deeply.
One of the ridiculous thing I hate the most about react is those hydration warning. It only exist because of shadow DOM syncing with real DOM.
Svelte doesn’t have that and manipulate DOM directly, and proved that we never needed shadow DOM. It even makes the app faster actually.
Enough to cover my rent and bills, but not much more.
After the release of V4 I needed to find a way to bring in more money, and seeing how many people asked how I made the https://t.co/jOWd0gBXWS landing page, I decided to make a course about that.
6/22
In 12 hours we've had 323 people sign our petition to protect local AI.
Open Source must win, not because anyone else must lose. But so that we can all win.
Please help me get 10,000 signatures so that we can walk into the room and say people care.
https://t.co/Z6WpNfSMXr
Totally agree with this.
You can actually start using these to accelerate your development.
But be wise and architect your application to easily replace them with self-host solutions or other providers without affecting your application.
"Just use Vercel."
"Just use Supabase."
"Just use Clerk."
Cool. Now your auth, database, and deployment are owned by 3 different companies who can change pricing whenever they want.
And the rest of your product is wrapping OpenAI.
At some point you have to ask yourself: what do I actually own here?
We've raised Claude Platform API rate limits for all users and simplified the tiers, which are no longer based on API spend.
The latest Sonnet and Haiku models now provide 5x higher rate limits at the highest tier.
FABLE 5 CAME BACK NERFED.
We re-ran the July 1st version of Claude Fable 5 on BridgeBench.
The results are brutal:
Debugging: 86.2 → 25.9
Refactoring: 73.6 → 38.4
Hallucination: 75.9 → 61.7
The new guardrails are kicking in on way too many tasks and falling back to Opus 4.8.
This is not the model that got banned.
Anthropic owes everyone an explanation.
Asimov 1 is an open-source humanoid robot you can build and customize yourself.
Two ways to get one:
1) Source the parts yourself: https://t.co/vtG89UlhiK
2) Get the DIY kit: https://t.co/tzvzNyXiq2
The kit bundles every part as a group buy, cheaper than sourcing one by one, and you build alongside others.
Karpathy LLM Wiki builds 56 linked pages in one hour with Obsidian and Claude, while NotebookLM forgets everything.
NotebookLM, ChatGPT uploads and every RAG tool share the same flaw. You ask a question, it re-reads your raw files, answers, then forgets. Tomorrow it starts from zero and nothing accumulates.
Karpathy flipped it. Instead of reading raw docs on every query, Claude compiles them once into linked markdown pages where concepts are defined, sources traced and everything cross-linked. Obsidian is the IDE, Claude is the programmer, the wiki is the codebase.
One writer loaded every essay he'd ever published and had 56 interconnected pages within an hour. Now it's the base he drafts from.
NotebookLM rediscovers your knowledge on every question. The wiki remembers.
https://t.co/f2ItnaCyLc
1/2
Eventually, even digital data will no longer be owned by individuals on their own initiative. Whenever there is a major change or accident in the world, in a country, in a government, in an idea, in a trend, access to it may suddenly be cut off.
if you’re building an app
make sure to apply to Apple’s Small Business Program
doesn’t matter if you’ve launched yet or not
it’s completely free
and cuts Apple’s fee from 30% → 15%
one of the reasons OpenCode 2.0 took so long was we redesigned it for hotreloading
if you ask it to make a skill for itself (or make one manually) it'll get picked up immediately in a way that does not bust cache
targeting public beta end of week, wish us luck!