tjkrusinski

The article is published primarily to signal to the market that Meta is serious in its efforts to compete in building frontier ai models.

They want to 1) attract talent, 2) tell wall street they can play in this space as well, 3) help employees feel the company is moving in the right direction.

A frontier LLM doesn't apply to their core consumer products.

Video goes over how to chunk code using tree-sitter and loading those chunks into Chroma for search and retrieval.

Are you constantly accelerating quickly? Haven't had this experience.

Interesting report. Are there recommended sizes for different models? How do I know what works or doesn't for my use case?

Zig is great, played with it a bit to compile to WASM and found it to be pretty easy to work with.

That said, they're not winning on the docs, community and marketing front vs Rust. It's not really apples to apples, but it's a comparison people naturally draw.

Looking back, the plan box does show that it is a free trial, but it's not clear at first as my eyes skipped the text that was in the parenthesis.

Love the idea. Since this is a consumer product, making it clear you can try the product before buying is critical. I clicked on the paid plan option, created an account but then was surprised to see that I ended up in the product.

I'd consider a react framework for the UI, shadcn or similar. The overall vibe of the UI doesn't present as cutting edge or on par with other AI based products out there.

In the product, I'd bring the recommended prompts to the surface the user enters the AI prompting mode, rather than hiding them behind the "show prompt tips" button.

The line height on the diff view is a little short, the text is crowded.

Tooltips on the toolbar buttons would be very helpful. Also, creating some hierarchy in the toolbar by dividing into logical sections grouped by tool type would help. Font foratting, node types, etc.

Overall, great job, product itself works really well.

(I have ADHD). First, exercise and diet are critically important, it's the foundation for the rest of your wellbeing.

A big thing for me is preparing my space to be able to get into hyper-focus mode. Do the dishes, clean the house, have everything else that can rob you of your focus out of the way.

You can take this same approach leading up to an interview. An hour or so beforehand, how can you get the flywheel moving to be in "productive" mode by the time the interview starts? Think about a few options here.

Action yields action, so getting the momentum going and keeping it moving helps you feel accomplished. When you feel accomplished, you feel better about yourself and you introduce a positive feedback loop.

You got this.

Looks good, neat idea.

Few things:

- Highlighting to then prompt makes sense, but I want to just click on the verse number to have it highlight to then summarize or get commentary on - The text to speech is useless, the system voice on chrome on macos is unlistenable compared to other modern speech synthesis models - The text formatting is strange, I'm used to bible.com or etc where formatting is preserved, not 'one line per verse' as you have it.

The .coffee domain is strange. Product looks good overall. I'm not sure why I'd want this over figma though. What's the difference? It wasn't clear reading the home page.