13
points
hooloovoo_zoo
Joined 931 karma
- hooloovoo_zoo parentHuh thanks that works but that’s not how silent mode is supposed to function. Still a bug imo.
- Don’t seem to get any sound on iPhone 16 even after clicking :(
- > Why do you think that the 2024 Putnam programs that they used to test were in the training data?
They reference https://artofproblemsolving.com/community/c13_contest_collec... for the source of their scrape and the Putnam problems are on that page under 'Undergraduate Contests'.
- For one thing, it's not a real score; they judged the results themselves and Putnam judges are notoriously tough. There was not a single 8 on the problem they claim partial credit for (or any partial credit above a 2) amongst the top 500 humans. https://kskedlaya.org/putnam-archive/putnam2024stats.html.
For another thing, the 2024 Putnam problems are in their RL data.
Also, it's very unclear how these competitions consisting of problems designed to have clear-cut answers and be solved by (well-prepared) humans in an hour will translate to anything else.
- These $ figures based on compute credits or the investor's own hardware seem pretty sketchy.
- My read is that they are describing functionality for site owners to provide input about what the site owner thinks should happen. OpenAI is not promising that is what WILL happen, even in the narrow context of that specific bot.
- Your link does not say they will obey it.
- It's just a new era in the ad-block wars where my ad-block LLM tries to excise the ads from their ad-generating LLM.
- Meta already bought all openAI’s secrets; now it just needs to let the GPUs cook.
- Maybe popsicle stick funny 'It was burnt out with being a rock-star'.
- 18 points
- Yes but you can also instantly advance to the state of that person’s former company’s research state which cost them way more than 250M.
- The had 100 candidates and hired him. Top 1% QED. (/s)
- Poor Sam Altman, 300B worth of trade secrets bought out from under him for a paltry few hundred million.
- Suppose that arXiv withdraws it and says the reason is fraud. What if it turns out to not be fraud? Either way, what if the author sues for libel? Why should arXiv spend resources evaluating papers after they've already been published on the arXiv? It's just inviting all the issues stackoverflow and the youtube copyright strike system have.
- Judging quality/fraud is the role of a journal/conference, not arXiv. If a paper gets rejected does it come off arXiv? No. If a paper is never submitted does it come off? No. If a paper is retracted, does it come off? No. ArXiv should avoid making as many subjective determinations as possible.
- I don't think arXiv should take it down even if it is fraud. ArXiv is more about being a permanent store than a quality judge.
- Plus a dumbbell is the same weight the whole time while the bow is only the draw weight at full draw.
- It looks very slow in the videos though.
- That study shows nothing of the sort. It essentially showed ChatGPT is better at pumping out boilerplate than humans. Here are the tasks: https://www.science.org/action/downloadSupplement?doi=10.112...
- iMessage is not a business; it's just a messaging feature. There are no ads or 3rd party content or anything. The easy litmus test is the one I just gave; users generally don't use their real names on tiktok.
- It's a good strategy because that's the obvious distinction and there's an easy litmus test (which apps do people use their real names on). Don't be ridiculous with iMessage.
- People have been gaming ML benchmarks as long as there have been ML benchmarks. That's why it's better to see if other researchers are incorporating a technique into their actual models rather than 'is this paper the bold entry in a benchmark table'. But it takes longer.
- It’s actually quite amusing as LTT used to have viewers bookmark their Amazon affiliate link in place of Amazon.com. Live by the sword…
- Is your prompt {$codebase} find bugs?
- It doesn't matter. They're all thin layers on functionally equivalent models. Stick with whatever text editor you prefer.
- It's not elitist. The Nvidia 'pro' cards (quadro etc.) have always been a slightly unlocked, wildly more expensive version of the consumer cards. The v100, a100, h100 are meaningful hardware upgrades to the consumer line.
- It still has to compete with renting an actual professional card(s).