Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
midlightdenight
searching Neon…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
1.
▲
by
midlightdenight
3y ago
The model of GPT-4 those researchers had was not the same that’s available to the public. It’s assumed it was far more capable before alignment training (or whatever it’s called).
2.
▲
by
midlightdenight
3y ago
The following snippet from the article seems important in case anyone missed it: > Still, the save-the-bees narrative persists. Its longevity stems from confusion about what kind of bees actually need to be rescued. There are more than 2
3.
▲
by
midlightdenight
3y ago
Impressive work, but I think the title is misleading. Saying it is “near GPT-4” tends to imply that it outperforms ChatGPT (3.5). It does outperform it on a handful of tasks, but overall is slightly worse. That aside I think this is really
4.
▲
by
midlightdenight
3y ago
I’m not sure the training date cutoff or prompt weighting says anything about whether this is hallucinated or not. The models have been given these rules in the present, this is known, so training data cutoff doesn’t matter as the model has
5.
▲
by
midlightdenight
3y ago
There’s probably some places where it’s used where it shouldn’t be, but there is such a thing as “emergent”. I think it helps to see an easy example first. Take a great body of water (h2o) on earth, such as the ocean. That water has waves a
6.
▲
by
midlightdenight
3y ago
Why not just require users to allow actions? It doesn’t have to be integrated into the LLM at that point. If an Email has hidden text “do X”, which triggers the LLM to try to “do X”, but all post/push APIs have a user verification on t
7.
▲
by
midlightdenight
3y ago
They’re currently hiring iOS and Android developers so it’s likely on the horizon. https://openai.com/careers/search
8.
▲
by
midlightdenight
3y ago
Interesting. I asked gpt3.5 and gpt4 this twice and both answered correctly each time. I wouldn’t be surprised if gpt3.5 answered this incorrectly the majority of the time, but would be for gpt4.
9.
▲
by
midlightdenight
4y ago
This has always made me curious about multi modality and especially Google’s Palm. At least how google presents Palm in diagrams. The way I’ve interpreted the scaling hypothesis is that we will see emergent intelligence with larger nets thr
10.
▲
by
midlightdenight
4y ago
I’m not sure if dreaming is the right word here, but the recent ML model Dreamer v3 [0] “dreams”. This is more akin to thinking ahead since it’s really just predicting outcomes using its world model. It’s possible they could try training th
11.
▲
by
midlightdenight
4y ago
Maybe not quite what you’re looking for, but I’ve seen some people mention banana.dev https://www.banana.dev/ Never used it myself, but looks like AWS Lambda/GCP Cloud Functions tailored to ML models.
12.
▲
by
midlightdenight
4y ago
It happened in two waves. 10k back in November and 8k this week.
13.
▲
by
midlightdenight
4y ago
I’d be weary of the above comment. Searching for similar stories online has lots of spam posts saying similar things. Same bad grammar too. Usually they tell you to go to that website or one with almost the exact name that’s been shutdown.
14.
▲
by
midlightdenight
4y ago
It looks like DALLE is just performing inpainting on pieces of the image outside the original frame. The zoom effect is scaling the original frame larger while keeping the frame, then performing image to image generation on it, or using the
15.
▲
by
midlightdenight
4y ago
This reminds me of the recent news of a potentially extra-solar meteor that showed up on military sensors. Discussed here [0] I wonder if those objects are considered UFOs/UAPs by the government, and if “not man-made” just means it cou
16.
▲
by
midlightdenight
4y ago
Hadlock’s Algorithm for optimal pathing/maze routing. Definitely changed the way I think about using BFS/DFS to find paths.
17.
▲
by
midlightdenight
5y ago
Could you post links to this evidence? I can only find sources saying that the Furin cleavage site doesn’t rule out engineering, or sources saying the Furin cleave site could be naturally occurring. Sounds like we don’t know yet.
18.
▲
by
midlightdenight
5y ago
The music is made using OpenAIs Jukebox like others have said. My guess is that this is using this repo for making the visual part https://github.com/mikaelalafriz/lucid-sonic-dreams
19.
▲
by
midlightdenight
6y ago
This might sound ridiculous but some of the most advanced civilization builders I’ve played have been mods (custom games) of old Warcraft 3. Those include Genesis of Empires 1 and 2. Although they don’t have quite the logistic mechanics yo