andai

15 ms·

511.
by andai 2mo ago
You thought they were gonna double the model size again? Also it occurs to me that they're somewhat incentivized to downplay cyber risks after what happened last time...
512.
by andai 2mo ago
Not 100% sure what clear does, but starting a new session invalidates the cache*, whereas I assume clear only removes part of the context, so it should be cheaper and faster. * In theory the system prompt is always the same and should th
513.
by andai 2mo ago
Austerity on tap!
514.
by andai 2mo ago
Precious submission's titles are more informative: > Nullschool Earth – Weather forecast by supercomputers updated every three hours > Earth: A global map of wind, weather, and ocean conditions
515.
by andai 2mo ago
If everyone bought from Temu then the net benefit would be zero relative to just including it in the price right? In practice the value flows from people who don't buy cheap crap from China to people who do.
516.
by andai 2mo ago
My fridge has a will to keep the food at a certain temperature.
517.
by andai 2mo ago
Menu css looks a bit dodgy with the notification at the top: https://files.catbox.moe/d5ztxi.jpg
518.
by andai 2mo ago
And, as is tradition, are deprived of approximately all the value they produce!
519.
by andai 2mo ago
Article should probably explain what "Mixedbread Search" is.
520.
by andai 2mo ago
An inside joke that confuses your target audience and makes you sound like a joke, is probably not what you want as your brand.
521.
by andai 2mo ago
Careful what you wish for! https://xkcd.com/810/
522.
by andai 2mo ago
Wtf, why isn't this the default? It says it can use any model/provider (but recommends Gemma?) I must investigate. This is probably where 80% of my cognitive load comes from these days: the "language barrier". (How ironi
523.
by andai 2mo ago
New: Announcing... Sleeper Agents!
524.
by andai 2mo ago
Yeah I saw complaints about this under their twitter announcement, that they're giving discounts to the rich foreign customers and screwing their own people.
525.
by andai 2mo ago
Have you compared it to Luna? I was using Flash for small tasks before, then switched to Luna when they dropped the price. The benchmarks show that Luna is significantly faster, but I think those are very complex tasks for which you'd
526.
by andai 2mo ago
Temu is Temu brand AliExpress.
527.
by andai 2mo ago
They appear to be targeting users with underdeveloped prefrontal cortex. I wonder what the intention behind that might be.
528.
by andai 2mo ago
> A hypothesis on the not-so-distant future ... in which most of what we consume gets cheaper, more abundant, and noticeably worse Isn't that just true about most things over the past century or so? They left out "staggeringly
529.
by andai 2mo ago
I'm not 100% sure but I also heard China subsidizes shipping of packages, to boost their economy. I worked as a mailman a while back, most days there would be more packages in grey plastic wrap (from China) than actual mail. And you co
530.
by andai 2mo ago
Trump 2020! I hate puppies!
531.
by andai 2mo ago
Roughly in order: data from simulated environments, data from robotics, data from brain waves.
532.
by andai 2mo ago
That could also just be random fluctuations in quality of service. Some days it's super fast, some days super slow or I get constant errors.
533.
by andai 2mo ago
OpenAI was bragging the other day about how HLE on Cerebras only took 11 hours. But you're saying if they just spun up a separate cheapo VM for each task, it would have finished faster and cheaper?
534.
by andai 2mo ago
Did the Python script call a vision API? Either way that's pretty impressive.
535.
by andai 2mo ago
It sounds like ChatGPT wrote it, but I'm assuming they used the model itself.
536.
by andai 2mo ago
> but isn't the cost for such a scan getting lower by the week Not with Anthropic's models!
537.
by andai 2mo ago
Agree. To this I would add, many things are saturated even for smaller models, which tend to be much cheaper and faster. On many of my tests, there was no difference in the result between the smaller and bigger model, but there was a big di
538.
by andai 2mo ago
Which one of those takes time?
539.
by andai 2mo ago
Ha, maybe. Sample size was low but usually both models ended up preferring the same solution.
540.
by andai 2mo ago
Yeah I forgot to specify "everyone with infinite money" hahahah. Not sure how you monetize a harness either way though?

More ›