• 1 Post
  • 96 Comments
Joined 1 year ago
cake
Cake day: July 13th, 2023

help-circle








  • Oh for sure, it’s not perfect, and IMO this is where the current improvements and research are going. If you’re relying on a LLM to hit hundreds of endpoints with complex contracts it’s going to either hallucinate what it needs to do, or it’s going to call several and go down the wrong path. I would imagine that most systems do this in a very closed way anyway, and will only show you what they want to show you. Logically speaking, for questions like “should I wear a coat today” they’ll need a service to check the weather in your location, and a service to get information about the user and their location.


  • I work on LLM’s for a big tech company. The misinformation on Lemmy is at best slightly disingenuous, and at worst people parroting falsehoods without knowing the facts. For that reason, take everything (even what I say) with a huge pinch of salt.

    LLM’s do NOT just parrot back falsehoods, otherwise the “best” model would just be the “best” data in the best fit. The best way to think about a LLM is as a huge conductor of data AND guiding expert services. The content is derived from trained data, but it will also hit hundreds of different services to get context, find real-time info, disambiguate, etc. A huge part of LLM work is getting your models to basically say “this feels right, but I need to find out more to be correct”.

    With that said, I think you’re 100% right. Sadly, and I think I can speak for many companies here, knowing that you’re right is hard to get right, and LLM’s are probably right a lot in instances where the confidence in an answer is low. I would rather a LLM say “I can’t verify this, but here is my best guess” or “here’s a possible answer, let me go away and check”.



  • The Pixels are probably the best high-end phone, but today the selection available is all bad enough that your choice comes down to “what features can I lose?”

    Even the likes of OnePlus have been shit for years. A company that literally entered the market on releasing an affordable flagship with near-stock software. Their last great phone was the OnePlus 6, before they decided to start ditching features.

    I had assumed that more companies would enter the market and take over, but that hasn’t happened. You still end up with no choice, whether it’s a poor screen, an awful camera, no storage, removed ports/jacks, no NFC support, or stupid little features that no one would actually give a shit about.

    The strength of early Android was that you had flagship phones that had the best new features, and experimental releases that tried new things on a budget like barcode scanners, slide-out keyboards, a desktop OS, remote features, etc. This still exists, but you’re paying even more for the pleasure of testing something in the wild.

    IMO, the world could use a new mobile OS, and one grounded in reality.







  • Working in Ruby did 10x more to help me write clean code than reading Clean Code ever did.

    Many of the lessons drilled into me with Ruby (keep a consistent style, tests are cheap, keep your methods relatively small where possible, reduce nesting where possible) carry over nicely into other languages without needing to go through any OO bullshit.

    IMO, the best lesson around Clean Code is this: you’re not clever, write obvious code that works with as few tricks as possible.