bryan0

14 ms·

211.
by bryan0 2y ago
Heh I just weighed mine: 28g (warby Parker frames). I really don’t notice them at all after a full day
212.
by bryan0 2y ago
While still pretty clunky, I think I'd rather wear these for extended periods of time than Vision Pro glasses (100g vs 600g). This is the type of v1 minimalist design I was hoping Apple would go for. I assume Apple investigated this pa
213.
by bryan0 2y ago
The locks have keys which are stolen: https://www.scrippsnews.com/us-news/us-mail-theft-often-begi...
214.
Lynch Yacht Sinking Off Sicily Proves as Baffling as It Is Tragic(nytimes.com)
5 points by bryan0 2y ago|1 comments
215.
The search for the biggest shape in the universe [video](youtube.com)
7 points by bryan0 2y ago|0 comments
216.
by bryan0 2y ago
I think this comment from the thread sums it up: “When DigiCert has another incident (and while I have tremendous faith in Tim, it will happen), I would rather that they have Jeremy Rowley with his wisdom and scar tissue around to guide the
217.
by bryan0 2y ago
This was previously submitted but flagged for some reason: https://news.ycombinator.com/item?id=41144106
218.
by bryan0 2y ago
Yesterday’s post said the probability of this anomaly occurring is difficult to calculate and showed that it happened 0 times in 1 million simulations. This new post by Tao provides a calculation where the probability is about 1 in 100 mill
219.
by bryan0 2y ago
If you look at the percentage breakdown it’s much more modest, so it appears to be an indication of how much the participants were already spending on alcohol. Also if you filter by “high income” beverage spending appears to switch to marij
220.
by bryan0 2y ago
You might be interested in this paper which was talked about on HN last week: “Computational Life: How Well-formed, Self-replicating Programs Emerge from Simple Interaction” https://news.ycombinator.com/item?id=40820022
221.
by bryan0 2y ago
You shouldn’t be so surprised, university PR depts are notorious for clickbait like this. These types of articles are all over HN. They are just incentivized to get their name in the press by whatever means necessary.
222.
by bryan0 2y ago
Discussed last month: https://news.ycombinator.com/item?id=40571995
223.
by bryan0 2y ago
Aaronson’s post on BB(5): https://scottaaronson.blog/?p=8088
224.
by bryan0 2y ago
Are there groups actively attempting this? I was thinking that this type of data gathering is a simple solution to LLMs “running out of data” to train on. Just put a camera in a room and let the model learn by exploring its environment.
225.
by bryan0 2y ago
2012
226.
by bryan0 2y ago
Like others I ran this through gpt-4o and it fails at reasoning because it does not realize Alice is a sister. In fact in one explanation it defines the siblings as: M brothers, N sisters, 1 Alice (!!) I asked gpt “Is Alice a sister?” It an
227.
Savings app CEO says 85,000 accounts locked in fintech meltdown(cnbc.com)
7 points by bryan0 2y ago|2 comments
228.
by bryan0 2y ago
The point is that even though meta “conducted extensive red teaming exercises with external and internal experts to stress test the models” a simple attack like this is still possible.
229.
by bryan0 3y ago
I recommend the Werner Herzog’s documentary “Fireball” ( https://en.m.wikipedia.org/wiki/Fireball:_Visitors_from_Dark... ) which has a segment with Jon Larsen and his process for capturing images of micro meteorites
230.
Class-action suit against company scanning and selling CA license plate data(sfgate.com)
7 points by bryan0 3y ago|0 comments
231.
Moon Lander Is Lying on Its Side but Still Functional, Officials Say(nytimes.com)
5 points by bryan0 3y ago|0 comments
232.
by bryan0 3y ago
Not if you include expected future legal fees. This also seems to imply if you train on Reddit data without a license Reddit will sue you.
233.
by bryan0 3y ago
I remember that mythbusters episode and I thought the conclusion was that the toothbrush bacteria cultures were no different than the control, which was a toothbrush right out of the package?
234.
by bryan0 3y ago
Even without custom domain blocking enabled, Kagi results are still somehow much much better (at least for my use cases). So clearly it’s not required to fix the seo spam problem.
235.
by bryan0 3y ago
This is really cool, I'm surprised I never heard of this project before. The examples look really clean. Most RAG tools seem to start with the LLM and add Vector building and retrieval around it, while this tool seems like it started w
236.
by bryan0 3y ago
I think people are way overthinking this. Guessing at random from the solution space converges surprisingly quickly.
237.
by bryan0 3y ago
Here’s some of the evidence of data manipulation: https://stanforddaily.com/2022/11/29/stanford-presidents-res...
238.
13-year-old mystery of Tweed Regional Museum's red letters solved(abc.net.au)
1 points by bryan0 3y ago|1 comments
239.
by bryan0 3y ago
You’re right. No AI has passed a reasonable Turing test, where both a computer and a human are trying to convince a judge they’re human over some extended period of time.
240.
by bryan0 3y ago
Gpt-4 fails the Turing test by default because it does not claim to be human therefore it’s trivial for a judge to identify it. It would be interesting to train gpt to pass the Turing test, but I think for any reasonable judge it would stil

More ›