dartos

34 ms·

511.
by dartos 2y ago
To your first point, the LLM still can’t know what it doesn’t know. Just like you can’t google for a movie if you don’t know the genre, any scenes, or any actors in it, and AI can’t build its own context if it didn’t have good enough contex
512.
by dartos 2y ago
> "Superintelligence isn't about discovering new things; it's about discovering new ways to discover." Wow I love that quote.
513.
by dartos 2y ago
It just depends when you want to draw the line of “this is beginning of a new culture”
514.
by dartos 2y ago
LLMs are trained with a data which may contain both truthful and false information. But at inference time it’s not referring to that data at all. Some of the data is aliased and encoded in the model’s weights, but we’re not sure exactly wha
515.
by dartos 2y ago
Why even have a button?
516.
by dartos 2y ago
Seems par for the course nowadays
517.
by dartos 2y ago
I think a lot of CEO were sales people. Being in that field, they can see where LLMs would fail at it. They can understand that you can’t just replace a sales team with LLMs. They don’t have that for engineering…
518.
by dartos 2y ago
I feel like every few years someone says we beat moore’s law. It wasn’t really a “law” in the first place, right?
519.
by dartos 2y ago
Hey I mean at least they self regulate at all.
520.
by dartos 2y ago
Idk YouTube comments are still a cesspit, but now there are lobotomized bots saying nice things.
521.
by dartos 2y ago
> one (or many) correct solution(s). > Code is a great example I’d argue that most coding problems have one truly correct solution and many many many half correct solutions. I personally have not found AI coding assistance very helpfu
522.
by dartos 2y ago
It’s probably a huge liability to not have very advanced and compliant crawlers. Accidentally ddosing several businesses seems like an expensive lawsuit.
523.
by dartos 2y ago
All I’m saying is that there is a huge machine backed my more money than can be reasoned about which doesn’t care about the environment. It’s been an almost 100 year battle to get laws and business culture to be more mindful. It’s clear tha
524.
by dartos 2y ago
I think the content is interesting, but anthropomorphizing AI always rubs me the wrong way and ends up sounding like marketing. Are you trying to market a product?
525.
by dartos 2y ago
That’s really great for you. I have been mislead several times in very subtle ways when trying to ask specific questions to chatgpt and Claude. It actually burned me at work. Made me look bad and caused my project to take an extra a sprint
526.
by dartos 2y ago
I think that attitude is a result of the tech startup culture. Companies get founded and funded with the exclusive goal of being bought up by a bigger fish. Why spend time investing in the company’s long term future when the goal is to not
527.
by dartos 2y ago
Sometimes it’s better to treat the symptoms if the root cause is out of reach.
528.
by dartos 2y ago
If we lived in a perfect world, we wouldn’t have let it get this bad in the first place.
529.
by dartos 2y ago
I think you’re putting people who make art on a pedestal. Remember there was an AI image that won an art contest. > To grandma it's basically an extension of gifs and emojis and memes, silly little novelty distractions Yeah of cours
530.
by dartos 2y ago
They don’t have to, but they can and it’s easy. They already have to spend the money on AI anyway to keep up in the stock market. It’d be a way to realize some return on their investment into GPUs.
531.
by dartos 2y ago
Pretty sure tech stocks boomed during covid.
532.
by dartos 2y ago
No, there are people on Facebook really do like AI slop. My Nonna unironically loves the AI Jesus gifs and AI gifs of dogs washing babies. Many elderly people, I think, don’t know or care if an image is AI generated.
533.
by dartos 2y ago
I’ll explain, but keep in mind that this is an exercise in cynicism, since none of us really know what’s going on. Currently many tech company stock prices are driven by their potential future with regards to AI. Investors are looking for A
534.
by dartos 2y ago
This one crazy hack is the secret to countless numbers of fruitful careers.
535.
by dartos 2y ago
Yeah. I’m a software engineer, and don’t find much use for LLMs for coding. But when I worked at a big (4k+ employees) org, the RAG service we used was very helpful. Everything from surfacing documentation, to announcements on confluence, t
536.
by dartos 2y ago
I think just about everyone company uses machine learning (read: statistical analysis) somewhere tbf. I think colloquially, when people say AI they mean LLMs.
537.
by dartos 2y ago
Is grinding code at a FAANG really that different?
538.
by dartos 2y ago
For those of you with a functional bend, check out elixir and phoenix. Elixir very similar to ruby in syntax and Phoenix very similar to rails in functionality. The creator of elixir (Jose Valim) was a core rails contributor back in the day
539.
by dartos 2y ago
Personally I see a lot of benchmarks come and go with the scores on these benchmarks always creeping upwards, but my practical day to day experience of using LLMs has remained pretty lukewarm. The things they were good at a year ago they’re
540.
by dartos 2y ago
I’d bet the same way any office worker gains value from LLMs. Summarizing things they don’t want to read. Generating things they don’t want to write, And RAG tools for surfacing documents, which imo is the killer app for LLMs.

More ›