thorum

16 ms·

181.
by thorum 2y ago
o1’s innovation is not Chain-of-Thought. It’s teaching the model to do CoT well (from massive amounts of human feedback) instead of just pretending to. You’ll never get o1 performance just from prompt engineering.
182.
by thorum 2y ago
If the rumors about the upcoming Strawberry and Orion models from OpenAI are true - supposedly capable of deep research, reasoning and math - they probably don’t have much to worry about. Not to mention they still have the only fully multim
183.
by thorum 2y ago
Is 84 so high? I imagine there are people sending party invites and “lost my phone, new number” messages to more than that.
184.
by thorum 2y ago
It probably has more to do with reputation & appearances than with specific practices. Meta is well known for its bad practices regarding privacy, OpenAI is not (yet).
185.
by thorum 2y ago
For adults, a better solution is to give users more active controls over the feed and its addictive elements. For example, the ability to remove kinds of content (“not interested”), put hard limits on minutes per day (Screen Time), hide Lik
186.
by thorum 2y ago
I wonder does it work in real time or does it add latency after each sentence?
187.
by thorum 2y ago
The information was tucked inside a blog post about the new safety board apparently: https://openai.com/index/openai-board-forms-safety-and-secur...
188.
by thorum 2y ago
Wow - the most impressive thing about this is the control options. I’m not aware of any other TTS systems with the same balance of control, quality and language support. Looking forward to testing this out…
189.
by thorum 2y ago
There’s still a Sky option but the actual voice has been changed.
190.
by thorum 2y ago
I don’t understand why self-driving systems don’t have any kind of fallback protections. Like however hard full self driving is to get right, surely a hard rule for “don’t drive at full speed into solid objects, no matter how good an idea i
191.
by thorum 2y ago
You might be interested in how CEV, one framework proposed for superalignment, addresses that concern: https://en.wikipedia.org/wiki/Friendly_artificial_intelligen... > our coherent extrapolated volition is "ou
192.
by thorum 2y ago
CEV is one possible answer to this question that has been proposed. Wikipedia has a good short explanation here: https://en.wikipedia.org/wiki/Friendly_artificial_intelligen... And here is a more detailed explanation:
193.
by thorum 2y ago
The superalignment team was not focused on that kind of “safety” AFAIK. According to the blog post announcing the team, https://openai.com/index/introducing-superalignment/ > Superintelligence will be the most
194.
by thorum 2y ago
Extra respect is due to Jan Leike, then: https://x.com/janleike/status/1791498174659715494
195.
by thorum 2y ago
Maybe? My definition of an art style would be something distinct and recognizably different like say, Van Gogh, Picasso, synthwave, or the aesthetic of one of Wes Anderson’s movies. I’ve seen AI blend, remix and do slight variations on thes
196.
by thorum 2y ago
The needle in the haystack test gives a very limited view of the model’s actual long context capabilities. It’s mostly used because early models were terrible at it and it’s easy to test. In fact, most recent models now do pretty good at th
197.
by thorum 2y ago
Large context models offer some hope for this. Gemini had a demo where the model was able to accurately translate a (human) language it didn’t know by putting an entire dictionary and grammar guide into the prompt. So maybe in the future yo
198.
by thorum 2y ago
A creative prompt can do new things with existing art styles. It can’t make new art styles.
199.
by thorum 2y ago
Many artists now view the tech industry as a credible threat to their work and livelihood, because of AI. If you want them to buy your products, it’s probably a good idea to show some sensitivity to that concern.
200.
by thorum 2y ago
It didn’t just not appeal to some audiences. It actively alienated one of the primary audiences of their product. Of course that requires a response.
201.
Ruler: What's the Real Context Size of Your Long-Context Language Models?(github.com)
2 points by thorum 2y ago|0 comments
202.
by thorum 2y ago
Discussion from the community: https://meta.stackexchange.com/questions/399619/our-partners...
203.
by thorum 2y ago
It’s an impressive model, but why would OpenAI need to do that?
204.
by thorum 2y ago
The text says “in a way that would be significantly more difficult to cause without access to a covered model” and in another place mentions “damage by an artificial intelligence model that autonomously engages in conduct that would violate
205.
by thorum 2y ago
The bill only applies to new models which meet these criteria: (1) The artificial intelligence model was trained using a quantity of computing power greater than 10^26 integer or floating-point operations. (2) The artificial intelligence mo
206.
by thorum 2y ago
One form this takes is that, even when you forget the details, you retain the general shape of the subject matter. You might not remember all the details about XYZ, but now at least you know XYZ exists: your internal map of the world is exp
207.
by thorum 2y ago
The increase from 2019-2021 (+2 years) was higher than the increase from 2012-2019 (+7 years). I wouldn't want to make any specific claims about the cause, but clearly the pandemic or the response to the pandemic was involved somehow.
208.
by thorum 2y ago
I found the graphs in the original linked article pretty helpful in understanding the trend. https://cdn.jamanetwork.com/ama/content_public/journal/cardi...
209.
by thorum 2y ago
I wonder how much a truly high-quality book recommendation service that actually worked and people trusted would change the landscape of the book publishing industry. My own experience as an avid reader is that it's quite hard to find
210.
by thorum 2y ago
GPT-4 and Opus are better for complex/precision tasks, and Haiku is cheaper for everything else. Good 7B/8B models are still really useful but let’s not be hyperbolic.

More ›