modeless

35 ms·

751.
by modeless 1y ago
The reduction in hallucinations seems like potentially the biggest upgrade. If it reduces hallucinations by 75% or more over o3 and GPT-4o as the graphs claim, it will be a giant step forward. The inability to trust answers given by AI is t
752.
by modeless 1y ago
12 to 24 months is old in mouse years. And the article offers a plausible explanation for both why plaques could cause the disease and why clearing them alone might not fix it without lithium supplementation.
753.
by modeless 1y ago
Lithium orotate is available over the counter. People could try it today. > Since lithium has not yet been shown to be safe or effective in protecting against neurodegeneration in humans, Yankner emphasizes that people should not take li
754.
by modeless 1y ago
All it does is make the pupil smaller? Sure that will improve sharpness, but less light will reach the retina.
755.
by modeless 1y ago
Cool, what software?
756.
by modeless 1y ago
The 20B one fits.
757.
by modeless 1y ago
What's the best speed people have gotten on 4090s?
758.
by modeless 1y ago
> Both of them are comparable to OpenAI's largest models and even bench favorably to Deepseek and Opus So are/do the new OpenAI models, except they're much smaller.
759.
by modeless 1y ago
I really want to try coding with this at 2600 tokens/s (from Cerebras). Imagine generating thousands of lines of code as fast as you can prompt. If it doesn't work who cares, generate another thousand and try again! And at $.69&#x
760.
by modeless 1y ago
The benchmarks of the OpenAI models are comparable to the largest variants of other open models. The smaller variants of other open models are much worse.
761.
by modeless 1y ago
Nah, these are much smaller models than Qwen3 and GLM 4.5 with similar performance. Fewer parameters and fewer bits per parameter. They are much more impressive and will run on garden variety gaming PCs at more than usable speed. I can&#x27
762.
by modeless 1y ago
Even better, it's 4 bit
763.
by modeless 1y ago
Can't wait to see third party benchmarks. The ones in the blog post are quite sparse and it doesn't seem possible to fully compare to other open models yet. But the few numbers available seem to suggest that this release will make
764.
by modeless 1y ago
Consistency over multiple minutes and it runs in real time at 720p? I did not expect world models to be this good yet. > Genie 3’s consistency is an emergent capability So this just happened from scaling the model, rather than being a
765.
by modeless 1y ago
I think AI will probably be capable of solving loneliness in an actually beneficial way: by encouraging us to seek socialization with other humans and by helping us improve ourselves to be better companions. Finally reversing the modern t
766.
by modeless 1y ago
OpenAI Services Agreement: "Customer will not [...] use Output to develop artificial intelligence models that compete with OpenAI’s products and services" Live by the sword, die by the sword.
767.
by modeless 1y ago
What purpose does "deactivating" any serve?
768.
by modeless 1y ago
You can share almost any video on YouTube to the kids app. You use the YouTube app on your phone to share it, not the UI in the kids app.
769.
by modeless 1y ago
Agreed the UX is terrible.
770.
by modeless 1y ago
> If there's some form of content that is not on Youtube Kids that you want to whitelist, you're out of luck. This is false too. You can add almost any channel or any video on YouTube by using the YouTube app on your phone to &
771.
by modeless 1y ago
This is an argument for not applying parental controls to YouTube for teenagers, while the guy I was replying to is explicitly asking for parental controls for YouTube for teenagers. I think "teenager" is too broad to have a produ
772.
by modeless 1y ago
As far as I can see the paper doesn't mention doing this at all. Do you have a secret source or did you just make it up?
773.
by modeless 1y ago
> And once your kid hits a certain age, they do not want to be on the kids version. Who's in charge here, you or your kids? Sure, maybe you could imagine a teen YouTube product you might like more, but you can't say the whiteli
774.
by modeless 1y ago
An unfiltered web browser has stuff a lot worse than YouTube. That's on you if you give your kids access to that.
775.
by modeless 1y ago
Yeah the interface in the Kids app sucks. The way to do it is from your own phone. Use the "share" feature and choose "with kids". It takes a lot of taps to share a channel, that could be improved for sure. But if you sh
776.
by modeless 1y ago
Sure, a whitelist makes YouTube less useful to a teenager, but it's hardly "pointless". Even a few whitelisted educational videos and channels could have huge value. You can send videos and channels to your kids' whiteli
777.
by modeless 1y ago
Why is everyone saying this doesn't exist? It's right there on the linked page! It's called "Approved Content Only" and I assure you that it exists, it's a real feature, it works just like you want, I use it my
778.
by modeless 1y ago
No. There is a true whitelist mode that only shows content you choose. Kind of weird that there are so many comments here lamenting the lack of this feature when it actually exists just as requested.
779.
by modeless 1y ago
YouTube kids has a feature to only show whitelisted channels and videos. It's been there a few years now. You can share videos to your kids directly from the YouTube app.
780.
by modeless 1y ago
I don't think it's true that 120 Hz flicker is easily perceptible, absent motion. Flicker fusion is real. But motion is common, so that's why humans in practice can detect 120 Hz flicker (and find it annoying. Ban PWM taillig

More ›