griomnib

12 ms·

181.
by griomnib 2y ago
I love how he’s claiming bias due to his critic’s employer. As though working for Google has no conflicts? A company that is desperately hyping each and every “me too” AI development to juice the stock price? Jeff drank so much kool aid
182.
by griomnib 2y ago
As an American…why not both?
183.
by griomnib 2y ago
NVIDIA sure as hell is trying to evade the spirit of the sanctions. Seriously questioning the wisdom of that.
184.
by griomnib 2y ago
Please elucidate me on your opinions about women’s rights then: Reproductive rights? Paid maternity leave? Equal pay? Equal civil rights? Acknowledging domestic labor as vital to the economy and supporting it? It seems to me that people who
185.
by griomnib 2y ago
Don’t threaten us with a good time.
186.
by griomnib 2y ago
People were here, and Europeans showed up and killed nearly all of them, claimed all the land as their “manifest destiny”, and proceeded to subject the few original inhabitants they didn’t kill to lives of desperation. Call that whatever yo
187.
by griomnib 2y ago
There can only be one Main Character, and Trump doesn’t share the spotlight with anybody. Elon will fuck up and his money won’t save him from what comes after that.
188.
by griomnib 2y ago
I’m sure they have the official MAGATEL licenses ready to go for the FCC as well.
189.
by griomnib 2y ago
Just the guts of the simulation showing through at higher resolutions.
190.
by griomnib 2y ago
Exactly, lots of people “lose” their savings in Vegas every day.
191.
by griomnib 2y ago
I know somebody who worked, planned, and saved to spend their retirement traveling the world. It was their life ambition. A few years after retirement they got early onset Alzheimer’s.
192.
by griomnib 2y ago
Hey now, the morons somehow managed to remain organized enough to choose these authorities. And now we’ll all burn together. Don’t let your fellow citizen off the hook.
193.
by griomnib 2y ago
The primary threat model is data collected today via mass surveillance that is currently unbreakable will become breakable. There are already new “quantum-proof” security mechanisms being developed for that reason.
194.
by griomnib 2y ago
I’ve literally spent 100s of hours with it. I’m mystified why so many people use the “you’re holding it wrong” explanation when somebody points out real limitations.
195.
by griomnib 2y ago
I agree with most of what you said, but “LLM can reason” is an insanely huge claim to make and most of the “evidence” so far is a mixture of corporate propaganda, “vibes”, and the like. I’ve yet to see anything close to the level of evide
196.
by griomnib 2y ago
Pretty much the same, I work on some fairly specific document retrieval and labeling problems. After some initial excitement I’ve landed on using LLM to help train smaller, more focused, models for specific tasks. Translation is a task I’ve
197.
by griomnib 2y ago
Anthropic had a vested interest in people thinking Claude is reasoning. However, in coding tasks I’ve been able to find it directly regurgitating Stack overflow answers (like literally a google search turns up the code). Giving coding is su
198.
by griomnib 2y ago
I felt the same way about a year ago, I’ve since changed my mind based on personal experience and new research.
199.
by griomnib 2y ago
Mathematical reasoning is the most obvious area where it breaks down. This paper does an excellent job of proving this point with some elegant examples: https://arxiv.org/pdf/2410.05229
200.
by griomnib 2y ago
Take a word problem for example. A child will be told the first step is to translate the problem from human language to mathematical notation (symbolic representation), then solve the math (logic). A human doesn’t use next token prediction
201.
by griomnib 2y ago
Likewise with LLM you don’t know if it is truly in the “chess” branch of the statistical distribution or it is picking up something else entirely, like some arcane overlap of tokens. So much of the training data (eg common crawl, pile, Redd
202.
by griomnib 2y ago
I think at this point it’s very clear LLM aren’t achieving any form of “reasoning” as commonly understood. Among other factors it can be argued that true reasoning involves symbolic logic and abstractions, and LLM are next token predictors.
203.
by griomnib 2y ago
He is a moderate; which tells you all you need to know about contemporary Israel politics and ethical standards.
204.
by griomnib 2y ago
QUANTUM AI BLOCKCHAIN!
205.
by griomnib 2y ago
…and a green line by the GOOG ticker.
206.
by griomnib 2y ago
Quantum computing + AI is undoubtedly the hype singularity.
207.
by griomnib 2y ago
https://en.wikipedia.org/wiki/Keyhole_Markup_Language
208.
by griomnib 2y ago
Oh, policy has always been a blood sport. You need a taste for conflict to enter the arena in the first place. (Been working on tech policy for a very long time, with scars to prove it!)
209.
by griomnib 2y ago
I wouldn’t say those are mutually exclusive. All humans deserve respect, but a lack of respect for expertise - particularly scientific - is a major problem globally and why our “intelligent” species isn’t doing anything about climate change
210.
by griomnib 2y ago
A flame war written in unformatted plaintext on a public list serve actually makes me nostalgic for simpler times.

More ›