psytrx

Joined 5 years ago

Karma

Submissions

Member since

Mar 9

Recent submissions

Aider is in development for almost 2 years now. Publishing another tool that "offers significant improvements" will have to catch up first.

Having options is a good thing, and approaching the dev agent problem space from a different perspective will help with pushing ideas in other products as well.

Comment · 1 year ago

There's llms.txt [0], but it's not gaining much popularity.

My web framework of choice provides these [1], but they're not easily injected into the LLM context without much fuss. It would be a game changer if more LLM tools implemented them.

[0] https://llmstxt.org/ [1] https://svelte.dev/docs/llms

Comment · 2 years ago

I've had the same thoughts on those. Maybe it also allows them to gather data on whether/how people interact with a webpage itself, which may be an indicator for its quality.

Comment · 2 years ago

Have you seen https://text-to-cad.zoo.dev/ ?

But yes, there's a lot of potential and, in my uneducated opinion, low hanging fruits for 3D printing especially.

Comment · 2 years ago

In addition to the initial loading time noted by the other posters:

You may want to use the same inference engine or even the same LLM for multiple purposes in multiple applications.

Also, which is a huge factor in my opinion, is getting your machine, environment and OS into a state that can't run the models efficiently. It wasn't trivial to me. Putting all this complexity inside a container (and therefore "server") helps tremendously, a) in setting everything up initially and b) keeping up with the constant improvements and updates that are happening regularly.

Comment · 2 years ago

Yes, too much water on the beans can lead to water gathering on the burrs, which can lead to rust on the burrs.

I use the teaspoon method (just the handle), cover it in water and stir the beans a couple of times. Just a single data point, but I've done this for 2 years at ~2.5 shots a day with no sign of rust on the burrs. YMMV

Comment · 3 years ago

Hey folks! I'm a motivated, generalist full stack software developer that builds on strong, practical fundamentals. I'm always looking to learn something new and expand my knowledge. I write software for more than 2 decades now, and currently focus on sustainable and accessible web development.

I'd love to find a remote position with a positive impact on the world, and a place where I can build awesome stuff with awesome people. I'd welcome a part-time position as well, if possible.

  Location: EU, Germany (UTC+2)
  Remote: Yes, preferred
  Willing to relocate: No
  Technologies: Full Stack, Svelte & SvelteKit, TypeScript / JavaScript, Go, Rust, .NET, Serverless & Edge (primarily Cloudflare)
  Résumé/CV: https://fdumont.net/resume?token=46b88224280d42f489acb437e326adf6
  Email: [see résumé]

Comment · 3 years ago

I'd love to find a remote position with a positive impact on the world, and a place where I can build awesome stuff with awesome people. I'd welcome a part-time position as well, if possible.

  Location: EU, Germany (UTC+2)
  Remote: Yes, preferred
  Willing to relocate: No
  Technologies: Full Stack, Svelte & SvelteKit, TypeScript / JavaScript, Go, Rust, .NET, Serverless & Edge (primarily Cloudflare)
  Résumé/CV: https://fdumont.net/resume?token=708e4b5d03524e88a201d9ab08e6c759
  Email: [see résumé]

Comment · 3 years ago

> I think it's eminently readable

This. And proper line breaks help a lot, too. There's a reason single line breaks are ignored.

> People have mentioned tables here, and...yeah, they're messy, ...

IIRC, the "original" markdown does not include tables. They've been added by GitHub (I think), because it made sense for them, and many parsers started to adopt. I agree they are not easily readable in "source mode". It helps to properly align stuff, though, but it's a PITA.

Comment · 3 years ago

These lanes and pull outs aren't common everywhere. When I visited NZ, it took me a day or two to realize they had a purpose. RVers usually are not driving in their home country.

I just recently tried to use the side lane in DK to let others pass (set signal, decelerate, pull over as far as possible). It actually confused the other drivers so much that they didn't dare to take over, even when there was lots of space to do so.

Comment · 3 years ago

I recently listened to a conversation between Lex Fridman and Stephen Wolfram [1]. Stephen said something that stuck with me for a while. Something along the lines:

GPTs are statistical word guessing machines. They regurgitate the average of the internet. You can get around that by placing with temperature and top-k, but that'll only make them less accurate, and increase the chances for them to hallucinate.

You, however, are an expert. Better than the average of the internet. Years of experience, and likely with higher standards. You understand the requirements, architect maintainable systems, and align with the company goals. You learned how to read requirements from context and between the lines.

Yes, GPTs can do highly complex things, and sometimes it feels magical or even scary. But you'll have to prime them to do so, and you'll have to guide them. They are your junior programmer, and you are their lead. You understand context, and they don't (simply because they don't have any).

Start to accept that from now on you have your personal junior dev. Learn to guide him into the right direction, and learn to provide accurate tasks. I'd also suggest looking into GitHub copilot chat to see where this is going.

If we're lucky, they'll get better with time and lots of research, but that's (likely) none of your concerns.

Uneducated opinion:

From this point on, I think it's likely the corpus they consume gets worse because they are starting to consume their own generated, average-of-the-internet content. You built your own Reddit bot, and know how easy it is to get generated content out there. It's a problem that isn't solved yet, and with time, it's becoming even more difficult because the number of different public models (which suffer from the same challenges) is steadily increasing, and therefore detection of such content gets a lot harder. The technology might get exponentially better, but in the end it boils down to the corpus they consume and learn from. This will slow things down, and will give you some time to learn to use these tools to your advantage, not see them as competition.

[1] https://lexfridman.com/stephen-wolfram-4/

Comment · 3 years ago

Very cool to see this indeed! I always had an eye on AstroNvim.

I'm currently on a slightly customized NvChad setup and am happy with it. It mainly transitioned me to lua configs, which has been a good decision.

How do you compare with NvChad? It's a slightly different set of tools, yes, but why did you think it needed another 'preconfigured vim' distribution?

Comment · 3 years ago

They definitely have sensible backout mechanisms for their free tiers (i.e. Pages, they still serve static assets but disable compute/Functions). Also, their docs are written pretty well, and openly hint at dangerous price traps.

We launched a big presale this year with an e-commerce app hosted on Workers, and we just barely hit the 5$ threshold for the paid plan. Did not use more services (other than DNS/proxy) though, so YMMV.

Comment · 3 years ago

Heck, 8k trees is a lot. Definitely find a professional. Growing coffee is exceptionally difficult.

Do you have an idea which crops have been planted? Is the plantation in a favorable location (altitude, weather)?

Depending on both, you should become clear about whether you want to produce commodity coffee or specialty coffee. The latter is more difficult to produce, but also more profitable.

If you're going the specialty route, it may even be fruitful (hah!) to get in contact with content creators. They usually have useful contacts in the industry, and they might be willing to connect you.

In any case, try to get in touch with European roasters, as they usually value traceability and prefer to roast single origin coffees.

Best of luck! And enjoy the journey.

Oh, and if you had your first batch, please drop me a message. I know a few local roasters (Germany) that may be interested in new sources. :-)

Comment · 4 years ago

It is pretty fast for me in central Europe. Maybe it's latency?