afro88

15 ms·

271.
by afro88 2y ago
Each question falls into a different category (ie math, coding, story writing etc). Typically models are better at some categories and worse at others. Saying "56% of people preferred responses from o3-mini" makes me wonder if tho
272.
by afro88 2y ago
Yeah. I immediately thought: I wonder if that 56% is in one or two categories and the rest are worse?
273.
by afro88 2y ago
The size reduction while keeping the model coherent is incredible. But I'm skeptical of how much effectiveness was retained. Flappy bird is well known and the kind of thing a non-reasoning model could het right. A better test would be
274.
by afro88 2y ago
But why is all that cereal $8? Where's the competition driving prices back down?
275.
by afro88 2y ago
Here's a better link: https://eqbench.com/creative_writing.html The R1 sample reads way better than anything else on the leaderboard to me. Quite a jump.
276.
by afro88 2y ago
> which I'm going to do as soon as I get my new personal website/blog set up I understand why you want a replacement for updating friends and family, but that's a really effortful barrier you're placing in front of de
277.
by afro88 2y ago
How accurate are the classifications?
278.
by afro88 2y ago
> This describes all websites everywhere You can't "follow" websites and get a uniform chronological feed of updates. That's partly why RSS exists, though it also doesn't give you a uniform feed of updates. That
279.
by afro88 2y ago
Another cherry-picked example of an LLM doing something amazing, written about with a heavy dose of anthropomorphism. It's easy to get LLMs to do seemingly amazing things. It's incredibly hard to build something where it does this
280.
by afro88 2y ago
Aider actually prompts the LLM to use search/replace blocks rather than actual diffs. And then has a bunch of regex, fuzzy search, indent fixing etc code to handle inconsistent respnses. Aider's author has a bunch of benchmarks an
281.
by afro88 2y ago
What we found was that error handling on the client side was also very important. There's a bunch of that in Aider too for inspiration. Fuzzy search, indent fixing, that kind of stuff. And also just to clarify, aider landed on search&#
282.
by afro88 2y ago
I'm interested in what stopped you from finishing diffs and diff based editing. I built an AI software engineering assistant at my last company and we got decent results with Aider's method (and prompts, and hidden conversation st
283.
by afro88 2y ago
Very interesting. Do you have any links with actual data / detail?
284.
by afro88 2y ago
I didn't say it, but I agree the rules are silly. However, they are what they are. What I was saying is that I think Magnus was protesting, and doing it as a publicity stunt. He didn't forget anything. He didn't expect to be
285.
by afro88 2y ago
This is silly. Surely a GM knows the rules. You can't say the facts are he wasn't doing it in protest or wasn't trying to make a scene. The facts are he violated the dress code rule and then withdrew when it was enforced. He
286.
by afro88 2y ago
Same
287.
by afro88 2y ago
Spotify isn't a monopoly, and if they want to fill their platform with stock music and presumably AI slop in the future, good luck to them. They're hollowing themselves out and making way for a new better service. And in the end,
288.
by afro88 2y ago
Not OP, but this isn't what Konfig did. They didn't try and rescue a bad idea by automating things. They attempted an AI focussed pivot. Using AI to automate things to rescue a failing product idea is kind of treating AI as a tool
289.
by afro88 2y ago
That's not an LLM, that's a subscription plan. You can select any OpenAI LLM on ChatGPT Pro. You can share the chat here, and this will show the LLM you had selected for the conversation. The initial prompt is also pretty importan
290.
by afro88 2y ago
Is o1 not one of the latest models? https://chatgpt.com/share/675ab91c-c158-8004-9dfc-ea176ba387... A better way to say it is: "funny how Google can't keep up with the reasoning effectiveness of OpenAI's
291.
by afro88 2y ago
Nope, o1. o1 "pro mode" is part of the $200/mo ChatGPT Pro
292.
by afro88 2y ago
How? 4o is part of the plus plan, as is o1.
293.
by afro88 2y ago
> but really saved me from moving to a different task and coming back You missed this part. Being able to quickly fix things without deep thought while in flow saves you from the slowdowns of context switching.
294.
by afro88 2y ago
Details of o1 Pro Mode here: https://openai.com/index/introducing-chatgpt-pro/
295.
by afro88 2y ago
If a script were applied that corrected "bad math" and now the LLM could solve complex math problems that you can't one-shot throw at a calculator, what would you call it?
296.
by afro88 2y ago
This is great, but my first question is: what is the infrastructure deployed? Is this a cool resume filler, or is it a serious alternative to Vercel? Seeing something around the infra and security in the readme would be great for production
297.
by afro88 2y ago
It's possible to plan for multiple eventualities. They may have pushed it to its limits (or beyond) and decided the best destination based on how it handled it
298.
by afro88 2y ago
They (YC) are interested in the use of LLMs to make the process of designing chips more efficient. Nowhere do they talk about LLMs actually designing chips. I don't know anything about chip design, but like any area in tech I'm ce
299.
by afro88 2y ago
> potential applications > if you ... > for example ... Yes there seems to be lots of potential. Yes we can brainstorm things that should work. Yes there is a lot of examples of incredible things in isolation. But it's a littl
300.
by afro88 2y ago
It's offending because it's complicit with the dystopian mass surveillance. Encouraging of it even. "I can find my keys when I lose them, so it's fine that a single corporation knows where billions of people are and what

More ›