Prosammer

I have very little understanding of geopolitics or economics, so these tariffs don't make sense to me, and they don't seem to make sense to most people.

What’s the best steelman argument for them?

I’ve read a bit of Peter Navarro and others who support this line of thinking, but I’m trying to understand: is there a coherent endgame here that benefits the country long-term, or is this just short-term political theater dressed up as strategy?

What would the best possible version of this policy look like if it were smart?

Check out lanceDB, it looks awesome for this.

Looks cool! Do you see this as something that should be handled by the MCP spec in the future? I'm surprised this isn't built in already.

I wasn’t suggesting anyone should duplicate code or that duplication doesn’t have downsides. My point was more about how LLMs, by default, tend to produce less “clever” (and often more repetitive) code. We know that duplicated code can be painful when bugs need fixing across multiple places, but I’m curious how programming practices will evolve moving forwards as I don't see this going away in the short-term.

>So if you're as clever as you can be when you wrote it, how will you ever debug it?

LLMs do not write very "clever" code by default. Without prompting them continuously to make it more "clever", they tend to write lots and lots of simple code, vs writing "clever" code that reduces redundant code, improves performance, etc.

What I am curious about is if these slop-filled codebases will be a problem or not in the future - traditionally it's been bad practice to have duplicate code everywhere, but with LLMs it feels like it matters less, as long as the code is simple and readable.

Caltrain runs reliably and safely for me, can you elaborate?

I'm a big fan of LLMs generally, but does anyone even want incoming texts summarized like this? Like even if they were accurate summaries, seeing "Wife expresses frustration with her husband's messiness" is a lot less fun than "Clean up your clothes, dipshit".

RepairWise (https://repairwise.pro) | Full-Stack Engineer | Remote, San Francisco, CA | Full-time

We’re a software-enabled startup on a mission to support the influx of thousands of off-warranty electric vehicles that need to be diagnosed and repaired, while delivering a trustworthy auto repair experience for EV owners.

The founding team has built and operated numerous well-known companies in the automotive & mobility space, and are now searching for exceptional full stack engineers to help pursue this mission. Experience with NextJS / TS / shadcn / tailwind / prisma is a plus!

Please email your Resume / portfolio to sam@repairwise.pro, and we’ll share more about the company and the role.

Very cool, congratulations! Would this work for graphql APIs with introspection disabled?

Yeah quite opinionated! I don't have an example project at the moment, but here's a youtube video from loco.rs that someone might find interesting: https://loco.rs/casts/007-htmx/

I've been playing with loco.rs to do this! Really cool, let's you use htmx or sveltekit (or whatever you want)

Looks cool! You should fix the "this is how our embedded demo looks like" grammar to say "This is what our embedded demo looks like"

I see what you're saying now, thanks for clarifying.

Sorry for the late response. I must be misunderstanding your comment. I read your comment as "RAG doesn't pre-compute KV for each document, which is inefficient". With RAG, you convert your text into vectors and then store them in a DB — this is the pre-compute. Then you just need to compute the vector of your query, and search for vector similarity. So it seems to me like RAG doesn't suffer from inefficiency you were saying it suffers from.

uh yeah it works out of the box, this is how most RAG systems are designed, just look at pgvector for example.

Why not both?

idk, if you think of it as two groups of people joining together to work together it's not so weird. people first vs business first maybe.

"Too agreeable" isn't an intrinsic trait of AI though, just the models you've been using. I'd love a less agreeable model :)

Personally, I've fallen in love with that visual effect of streaming text you're talking about. It's a bit pavlovian, but I think in my head it signifies that I'm reading something high signal (even though it isn't always).

Can you provide a source for this? This has not been my experience, nor have I heard this anywhere before, and I am quite involved in the psychedelic community.

I've been finding with these large context windows that context window length is no longer the bottleneck for me — the LLM will start to hallucinate / fail to find the stuff I want from the text long before I hit the context window limit.

Cheers!

Yes, I'm stoked you brought this up, as I had previously googled and found some articles that recommend doing this in typescript. I haven't tried it out yet though, mainly because it seems like every article has a different approach to production-ready error handling in typescript and I don't know the best approach. Do you / does anyone have any suggestions for articles etc. on this?

yeah honestly I built 1 little mac app in Rust and now I'm using typescript and can't stop thinking about how I miss Rust's Option and Result

Thanks for saying this! Really appreciate it :)

I am building textool [1] an app that lets you create endpoints using GPT4. The idea is to make it so you can create "actions" for GPT4 assistants easily.

  - Nextjs
  - Deno Deploy for hosting the apis 
  - Supabase - postgres / auth
  - Shadcn

I want to use the t3 app stack [2] for v2.

It's really MVP, but I want to see if anyone is interested at all before I work on v2: creating gpts that come with databases!

  [1] https://textool.dev
  [2] https://create.t3.gg/

SEEKING WORK | REMOTE | Canada Frontend: React, t3 app stack, Svelte,Tailwind, shadcn, Trpc, nextauth, prisma, zustand, sveltekit, next.js, sveltekit

Backend: Rust, Node, Electron, Tauri, PHP, Kotlin GCP, Firebase, Supabase

Full-stack dev who left fintech so I could work on my own SAAS. Open to flat-rate pricing for well-scoped projects.

I've got good social skills and can communicate clearly. I love shipping MVPs fast!

Contact at sam.finton [at] gmail.com or find my linkedin.

My understanding is that these vector search databases are generally used by people who want to use an existing AI and extend it with RAG etc. Was anyone ever expecting major AI players to use a tool like this as you are suggesting?

If I can't make my pitch / problem simple enough so that the average person / mom can understand it, I need to refine it. See: YCs advice on building simple pitch decks

I will look into these, thanks for the suggestions!