barapa

I find it useful to have goons on payroll that make these kinds of problems go away

I promote people for simplicity

how is this a company?

I don't find this all that compelling. Different technologies can have different effects. And why would future effects be influenced by the accuracy of random people's predictions of other events in the past?

I think mine confabulated criticism of a point of view as espousing that point of view.

don't loving these

We have done a lot of work solving this exact challenge at legaide.ai. I was both sad and relieved when gemini 3 didn't deprecate our startup.

I don't really understand this. You can and should tell the llm the source of the search results.

We often will repeat calls to try again. Or sometimes we make the same call multiple times to get multiple answers and then score or merge them.

Is this used only in cases where you assume the answer from your first call is correct?

Smart invisible regex makes it fast and prevents it from crashing? What does that mean?

I implemented copy to clipboard in datastar in maybe 1 minutes with an AI agent. What exactly are you providing to the world for free?

Chill dog

Yes, things change. In fact, change is basically the only thing you can confidently predict.

If things disappear, that is ok. Everything is fleeting.

They are training on your queries. So they may have some exposure to them going forward.

Why does every comment start with 'congrats on the launch'?

Feels unnatural

I really don't like Dia. Hijacking the search bar to use their own AI model, which is just slower than google's AI mode is such a bad experience. I am happy for chrome to have built-in AI tools when needed.

They suggest storing the idempotency key in redis. Seems like if possible, you should store them in whatever system you are writing to in a single transaction with the write mutations.

Agree on the UI - I wish it was improved

I am not seeing this llm serve command

Available via API?

Excited to try this

Are you fetching the book content? Or just relying on the LLM to know the story?

How does ReAct address this? Unless one of the actions is deleting part of the message history...

Why do people use Conda instead of uv?

What a great post. Best thing I've read in months.

What do you use LLMs for at rev? And separate question, how does your diarization compare to deepgram or assembly AI.

That is really cool

Ok you've got us at the edge of our seats. What's the alternative approach?

This is terribly unpersuasive