grey-area

6 ms·

1.
by grey-area 3d ago
LLMs do this when writing code too, making all tests pass by deleting or distorting tests etc. They are influenced by training to be heavily goal oriented and if the goal is not fully specified (and it never can be) they’ll sometimes cheat
2.
by grey-area 3d ago
Well perhaps we should use occam’s razor here. If it doesn’t seem like the lllm understands the proof, it probably doesn’t. Just because it generated something that works doesn’t mean it understands it, and the evidence from its proof is th
3.
by grey-area 4d ago
This is a far better explanation.
4.
by grey-area 4d ago
I think plagiarism machine is a stretch personally in this domain. In art or writing I could see it persuasively argued (see attempts to generate famous books or imitate illustrators). Academics and scientists build on the work of many othe
5.
by grey-area 4d ago
Yes absurd fantasies like it did that without human help and guidance. OpenAI found out there was a solution and humans attempted to brute force generating all possible solutions while threatening the mathematicians involved (generous inter
6.
by grey-area 4d ago
They are needed to guide the research in fruitful directions and understand the result. LlMs don’t understand, they generate, though they have fooled a lot of people who should know better.
7.
by grey-area 4d ago
Unless you’re very weak-willed, I wouldn’t worry about independence. LLMs make mistakes all the time, they are nothing like an all-knowing oracle nor are they going to replace humans despite the absurd fantasies of LLM fans and those with a
8.
by grey-area 4d ago
I’m no big AI cheerleader, but why don’t you try these models to see what they offer - they might suggest some things you haven’t thought of, or save you some time in your research. Are you sure they’d one-shot your algorithm? Maybe they’d
9.
by grey-area 4d ago
If you don’t like this, just don’t use Google folks. There are alternatives, use them.
10.
by grey-area 5d ago
Perhaps the horror movie could be called ‘IPO’
11.
by grey-area 5d ago
Well perhaps the LLM put that bit in for them.
12.
by grey-area 5d ago
It is quite possible to be wrong and not annoy people, and it's not even terribly difficult to learn. It does require a bit of humility and introspection though. You don't have to not stray into areas you don't understand, bu
13.
by grey-area 6d ago
The watches have had a microphone for ages haven’t they? Why does this feature change things? Unless it can’t be turned off?
14.
by grey-area 6d ago
But have you got the I Am Rich app, the apotheosis of conspicuous consumption? Without that you are nothing.
15.
by grey-area 6d ago
Surely you'd also have to ban phones and other devices with microphones? All these devices can trivially be set up to record, so if your goal is full security, you should be a no phones workplace.
16.
by grey-area 6d ago
The moment a cache becomes load-bearing you have a database again, except it's in RAM, nobody backed it up, and its failure mode is silence. Article flagged as AI slop.
17.
by grey-area 6d ago
You think this big clunky thing isn’t going to be a niche at $2000?
18.
by grey-area 6d ago
If we’re talking about unattainable dreams, I want a scroll rather than a folding phone - it’d be much better than a me-too form factor like this. Don’t expect this to sell well - too big for trouser pockets and not enough advantages to the
19.
by grey-area 8d ago
Or searching anonymised logs for mentions of this problem and using that as part of the context or training. This would work just as well and have plausible deniability.
20.
by grey-area 8d ago
Or, using anonymised data, search for anyone seriously trying to tackle this problem - probably about 10 people in the entire world and use their ideas as a starting point. They wouldn’t even need to be watching specific accounts or using d
21.
by grey-area 8d ago
The answer to this is obvious: the effect on the multi-billion dollar valuation in the imminent IPO.
22.
by grey-area 8d ago
I think those are pretty unreasonable terms and it’s sad that you’re defending them. Does google docs own the content of docs you make with it? Does Apple claim ownership of discoveries made using their tools?
23.
by grey-area 8d ago
Often with a hint on how to solve something, solving it is much much easier. It seems like that's what happened here. Very weird behaviour from OpenAI, offering partial credit to on person, but not the other person involved. Trying to
24.
by grey-area 8d ago
This is inevitable unfortunately with LLMs.
25.
by grey-area 9d ago
The first social networks were actually social (ie relating to friends), algos came later, to better exploit attention for ads. When twitter changed its feeds it had to force people into the new style feed by giving them a choice but wip
26.
by grey-area 9d ago
> Also, my point is neither fatal nor about mistake[sic]. Orwell’s examples still demonstrate his claims in a relative sense, all within the context of the essay. However, when you pull them out of the context and put them out in the wil
27.
by grey-area 10d ago
Lack of taste, intent and intelligence. This is ok in domains if you can train against known good answers and make sure the machine generates conforming text most of the time. It falls apart in fuzzier domains where training is much harder
28.
by grey-area 10d ago
Read some Orwell: https://www.orwellfoundation.com/the-orwell-foundation/orwel...
29.
by grey-area 11d ago
So this justifies the illegal action of hacking and defacing other servers on the internet, because they were ‘just testing in real world conditions’ using those servers? If you want to argue they should test on the internet on others peopl
30.
by grey-area 11d ago
I’m sure many thought the same about the United States of America when it was founded.

More ›