krackers

53 ms·

541.
by krackers 10mo ago
>It used to rightfully be something we looked forward to Science fiction has always been mixed. In Star Trek the cool technology and AGI like computer is accompanied by a post-scarcity society where fundamental needs are taken care of. T
542.
by krackers 10mo ago
This is almost surely in violation of a bunch of corporate policies... The whole point is that they don't want you having access to anything once access is cut.
543.
by krackers 10mo ago
> bed frames and box springs are mostly superfluous and a mattress on the ground is more comfortable and cheaper. This is basically a Japanese futon. The only con I can think of is the one the other commenter noted, about mold buildup in
544.
by krackers 10mo ago
I wonder if semi-reliable RAM could be made to work for training. After all gradient descent already works in a stochastic environment, so maybe the noise from a few flipped bits doesn't matter too much.
545.
by krackers 10mo ago
You can analyze this in various ways. At the "next token predictor" level of abstraction, LLMs learn to predict structure ("hallucinations" are just mimicking the style/structure but not the content), so at the stru
546.
by krackers 10mo ago
There's a history of things like this: pfas, tobacco, asbestos, BPA, which are later found to cause issues. Generally regulation seems to be under the principle of "no concern until dangerous side-effects found" when it shoul
547.
by krackers 10mo ago
I can believe this, Deepseek V3.2 shows that you can get close to "gpt-5" performance with a gpt-4 level base model just with sufficient post-training.
548.
by krackers 10mo ago
https://news.ycombinator.com/item?id=23514072 and https://news.ycombinator.com/item?id=12354494 seem to contradict this and claim that modern intel processors don't use separate AGU for LEA... Not too
549.
by krackers 11mo ago
Hm the ones I can find have a heavy green reflex and their optical density seems about twice or thrice what you'd need for a 1W CW laser. Maybe it's unavoidable given the closeness of near IR and deep red, but I wonder if there ex
550.
by krackers 11mo ago
If you find anything, please let me know. The least obtrusive option I've found is this Zeiss lens coating ("Thermo Force") which claims to block 90% of IR "between 780 and 2000 nm", which covers the NIR used in bot
551.
by krackers 11mo ago
The visual system can patch over tiny defects (see: blindspot) and visual field tests have not been part of standard yearly eye exams I've been to. And possible longer-term risks (say increased risk of cataracts) would be harder to con
552.
by krackers 11mo ago
My favorite part about the spread of AI/LLM stuff is that it opens up a new kind of reverse engineering. Trying to fetch the system prompt that was used. Trying to deduce the model that was used (there's lots of ways to do this: g
553.
by krackers 11mo ago
I was always curious about this, it's impossible to find any safety certifications or details about the lidars used by e.g. Waymo. Are we supposed to just trust that they didn't cut corners, especially given the financial incentiv
554.
by krackers 11mo ago
This won't happen in tech companies because that would make legal discovery a lot easier.
555.
by krackers 11mo ago
If the error message happens to include a numeric code, OSStatus.com is sometimes helpful if the issue if they didn't bother with a localizable string for it.
556.
by krackers 11mo ago
>you would easily make any salary you could name for them tenfold. >I at at a Home Depot like 10 times a week And yet you still go to Home Depot, so from their perspective it's not an existential issue. Probably the biggest thing
557.
by krackers 11mo ago
The more interesting thing is that they were actually pulled. Did the manufacturer send a recall notice? (And if so why didn't they send a statement). I can't imagine Amazon themselves taking any action since they allow other junk
558.
by krackers 11mo ago
Doesn't hellen keller provide a counterexample? She seemed to imply pretty strongly that before acquisition of language she operated more on stimulus and bodily perception rather than higher-level thought.
559.
by krackers 11mo ago
We will need people to work, but not everyone will have the capacity to fulfill that work, as jobs become increasingly specialized and the floor goes up
560.
by krackers 11mo ago
The real world is governed by physics, so isn't "next state prediction" a sufficient eval function that forces it to internalize a world model? And increasing the timespan over which you predict requires an increasing amount
561.
by krackers 11mo ago
>What high quality data sources are not already tapped Stick a microphone and camera outside on a robot and you can get unlimited data of perfect quality (because it by definition is the real world, not synthetic). Maybe the "AGI ne
562.
by krackers 11mo ago
There was some HN article a month back about how all types of gambling: sports gambling, prediction markets, etc. are becoming more popular. All of this presumably downstream from a felt perception of economic malaise and nihilistic sense o
563.
by krackers 11mo ago
What triggered the bubble pop in this case?
564.
by krackers 11mo ago
I think in JS it's easier to see because of the correspondence between Promises and async/await. So in your example the behavior is much more obvious if you sort of desugar it as async function parent() { print("par
565.
by krackers 11mo ago
https://www.mit.edu/people/dpolicar/writing/prose/text/think...
566.
by krackers 11mo ago
Snow Leopard wasn't 0 new features, it brought pretty sweeping kernel level changes. Although I suppose being able to focus on low-level developer facing things rather than flashy new consumer facing things is still in line with the ph
567.
by krackers 11mo ago
>I don't blame them, but something must be off in how they were trained. No, you should blame them. To this day optometrists will prescribe children glasses despite there being an ample body of evidence that working at distances clo
568.
by krackers 11mo ago
I think long-term just having a single LLM responsible for everything will win out compared to brittle and complex subagent hierarchie. Most use of "subagents" today are just workarounds for LLM limitations: lack of instruction fo
569.
by krackers 11mo ago
I'd maybe make a hypothesis that a large portion of the space is "bad" smelling stuff: smoke or garbage. When people had covid-induced parosmia, it almost always seemed to be bad smelling stuff.
570.
by krackers 11mo ago
I'm guessing it's the fact that linux has in-tree drivers, so you necessarily need to "patch the kernel" in order to write/fix a driver for a non-standard compliant device?

More ›