DalasNoin

3 ms·

1.
Trust AI leaders when they AI is dangerous, but don't trust them when(simonlermen.substack.com)
3 points by DalasNoin 12d ago|0 comments
2.
by DalasNoin 12d ago
what you read (past tense) is already the corection
3.
by DalasNoin 13d ago
thank you for this reasonable reaction
4.
by DalasNoin 13d ago
"Every single one of these catastrophic breakouts happened inside the testing environments of the exact same vendor." This is incorrect, the HF incident for example (the most well known) had nothing to do with irregular. I know th
5.
by DalasNoin 14d ago
"Me: a poster paint drawing by his young child." Some time ago, I was in this apartment complex and there was a hand drawn sign of a dog pooping on the grass - the sign then said that residents should pick up after their dog. It l
6.
by DalasNoin 22d ago
Huggingface was attacked by models that finished training earlier this year, perhaps May. Current models are already substantially stronger. the next incident could be happening now. There is certainly no clear reason why models shouldn&#x2
7.
AI Voice Phishing Performs on Par with Human Scammers at a Fraction of the Cost(fredheiding.substack.com)
2 points by DalasNoin 3mo ago|0 comments
8.
Where does the race to automate AI research end?(simonlermen.substack.com)
2 points by DalasNoin 4mo ago|0 comments
9.
by DalasNoin 5mo ago
Can you confirm people in france actually use wero? I had heard of it every so often but basically zero people actually use it, my revolut app has a feature to use wero but never used it. I mean would be great, getting rid of CC fees could
10.
by DalasNoin 5mo ago
Quote from the article: ""AI will probably most likely lead to the end of the world, but in the meantime, there'll be great companies," Altman said in 2015." Altman wasn't even at OpenAI at that point, so why w
11.
by DalasNoin 5mo ago
I mean the best argument I see for cursor is that you can easily switch between AIs, which is convenient since they seem to run at 80-90% up time (with those 10-20% clustered at West coast working hours). But the big AI companies are likely
12.
by DalasNoin 7mo ago
Why does SynthID make it worthless? it helps other platforms detect this as ai?
13.
by DalasNoin 7mo ago
We use semantic information inferred from comments and submissions. I think using stylometry would be a great addition, but it would be hard to google for "guy who writes fanciful using many puns" rather then "indie developer
14.
by DalasNoin 7mo ago
We test different methods, in section 2, we use LLM agents to agentically identify people. We don't share any code here, but you could try with various freely available agents on yourself.
15.
by DalasNoin 7mo ago
We essentially don't use stylometry but semantic information revealed from peoples' comments – clues and interests. (We use a little stylometry in a single experiment in section 5)
16.
by DalasNoin 7mo ago
We essentially don't use stylometry but semantic information – clues and interests.
17.
by DalasNoin 7mo ago
That's a great background paper on the Netflix attack, we make a pretty direct comparison in section 5. We also try to use similar methods for comparison in sections 4 and 6. In section 5 we transform peoples Reddit comments into movie
18.
by DalasNoin 7mo ago
I agree that these accounts probably on average still contain more information than the average pseudonymous account. I think we could try to use the LLM to increasingly ablate more information and see how it performance decays – to be clea
19.
by DalasNoin 7mo ago
We do advocate for stricter controls on data access on social platforms because of this. There is a bit of an unfortunate trade-off, but I think allowing mass-scraping or downloads of data from social sites can be misused in increasingly mo
20.
by DalasNoin 7mo ago
Mitigations are pretty difficult, I understand it is kind of cool that some websites have really open APIs where you can just read everything. There are some cool apps that used HN data in the past. But I think there should at least be cons
21.
by DalasNoin 7mo ago
There is also a practical issue here that people usually don't write a lot on linkedin, most people just have structured biographical information. We use very limited stylometry in section 6 for matching reddit users who we synthetical
22.
by DalasNoin 7mo ago
We don't use (much) stylometry, so this won't help. This is totally something you could try, but we use interests and clues. Semantic information you reveal about yourself. The blog post might be more approachable if you want to g
23.
by DalasNoin 7mo ago
To be clear, we are making a clear concession here that the people weren't truly anonymous. But we did use an LLM to remove any identifying information from HN making them quasi-anonymous, this is more described in the appendix Table 2
24.
by DalasNoin 7mo ago
There goes all the prompt engineering jobs
25.
Large-Scale Online Deanonymization with LLMs(simonlermen.substack.com)
364 points by DalasNoin 7mo ago|234 comments
26.
Large-scale online deanonymization with LLMs(arxiv.org)
3 points by DalasNoin 8mo ago|0 comments
27.
by DalasNoin 8mo ago
From what I understand this is a fully open-source bot that anyone can run with no restrictions. what a time to be alive, let's see what these bots will break first
28.
by DalasNoin 8mo ago
The energy collected from the solar panels must be converted into heat in the AI chips. It's really like putting the AI chips directly into the sun, just with extra steps. Sunlight gets transformed into electricity which gets transform
29.
by DalasNoin 8mo ago
It's cold in a sense that is not very relevant. Your tumbler has a vaccuum layer because vaccuum does not transport or absorb any heat. you need those atoms to carry away heat.
30.
by DalasNoin 8mo ago
Space doesn't seem like a good place to build datacenters at all. Cooling is going to be an enormous issue, how do you disperse of heat in a vaccuum? Radiators are very ineffective for cooling.

More ›