foota

26 ms·

481.
by foota 2y ago
I wonder if they could somehow feed in a trained Gaussian splats model to this to get better images? Since the splats are specifically designed for rendering it seems like it would be an efficient way for the image model to learn the geomet
482.
by foota 2y ago
Price shocks in anything are generally bad for the economy (maybe with the exception of some commodities like oil, but even this can be bad since the producers of that commodity will be in trouble). You're too focused on the rich peopl
483.
by foota 2y ago
They're not concerned about the wealthy, but the state of the economy. Bad things happen when the prices of things change dramatically. E.g., if you happen to own an asset that a billionaire now needs to fire sale, you'll lose out
484.
by foota 2y ago
Maybe I misunderstand, but if the ring buffer is full isn't it ok for the sender to just block?
485.
by foota 2y ago
It seems like they're doing it differently than they did for e.g., China. Note that the money is simply a matter of a contract (e.g., we will hire your company, which is located in China to operate our cloud region. We'll give you
486.
by foota 2y ago
The operator isn't under the other company, so if they say "we need this data" they can just say no. Now potentially they could try to trick the operator, but I'm not sure a company could be compelled to do so under US l
487.
by foota 2y ago
These sovereign clouds generally put the root in trust with a local operator so they physically can't be compelled to release information.
488.
by foota 2y ago
Why would you prefer this when the alternative is no downtime and the provider forgiving the bill?
489.
by foota 2y ago
When they don't have to interface with JS apis (e.g., for the DOM) this is true for libraries in other languages built for WASM.
490.
by foota 2y ago
These issues don't exist in languages that aren't built on a marsh. More specifically though, I feel like the way javascript libraries with types work is often painful to use and that's why people use TS's escape hatches
491.
by foota 2y ago
This sort of "worst intentions" stuff is FUD. Sure, you can leak some potentially embarrassing photos of someone with doctored drugs on the floor. So what? I can Photoshop some powder into a picture too. It might look better, but
492.
by foota 2y ago
Imagine if we found a biological system doing Reed Solomon coding
493.
by foota 2y ago
Ah, interesting. I guess you could probably get bounding boxes etc.,. too, so you could pass the image as well to a multimodal model (although maybe it'll just run tesseract behind the scenes lol)
494.
by foota 2y ago
I wonder if you could feed back the results from an LLM into the OCR model to get it to make better decisions. E.g., if it's distinguishing a 1 from an I, the LLM could provide a probability distribution.
495.
by foota 2y ago
Is this just the location data being trained on, or is there image and sensor input data too? It looks like it's just location, which seems like it limits the applicability, but I'm not sur Edit: reading a bit more it's somew
496.
by foota 2y ago
I also really like rust's support for deriving traits. Being able to generate hashes, equality, debug strings, etc., is really convenient for hacking around.
497.
by foota 2y ago
It's sustainable in the short term (for a very long definitoon of short term) though, which is what really matters wrt migrating off of fossil fuels. We could pull many times humanity's energy consumption from the earth for a cent
498.
by foota 2y ago
I found https://github.com/embassy-rs/ekv after some searching, but who knows how battle tested it is.
499.
by foota 2y ago
A columnar database format might perform well out of the box in that case. Reading more, it seems like a columnular database built on LSM trees would probably work great, since you could easily move the log around in flash to distribute wri
500.
by foota 2y ago
Generally, yes. But implementing UBI by wasting people's time is a bad idea.
501.
by foota 2y ago
This seems strongly to me like a broken window argument. Taken another way, you're paying people that could likely be doing more productive things to sit in a car and burn gasoline. If money draining out of the economy is a major conce
502.
by foota 2y ago
It sounds like you're familiar with the field, so I'll accept that you're probably correct here, but the wikipedia article doesn't say what you're claiming. It says coulom explosions are made with ultra short lasers
503.
by foota 2y ago
These are just discussing femtosecond lasers, I don't see any mention of coulomb explosions? Maybe it's implied, but I don't see it.
504.
by foota 2y ago
The only references I can find online seem to be for studying materials (e.g., https://www.scienceinschool.org/article/2024/explosive-imagi... ) and for nano machining (e.g., https://www.nature.com&#x2F
505.
by foota 2y ago
How about this: take as input something random from beyond the time horizon (e.g., brightness from stars 300 light years away). Then (if we assume the input is sufficiently random) they should be fine.
506.
by foota 2y ago
I don't see the malicious part of the compliance here. Maybe lazy compliance?
507.
by foota 2y ago
I've dabbled in trying to do procedural generation in the past (in particular to make good looking trees), and I feel like the missing component (to me) was easy ways to connect geometry. It's easy to make two cylinders, but then
508.
by foota 2y ago
Never? Edit: To defend my response, the model definitely knows when it hasn't yet found a correct response, but this is categorically different from knowing that it does not know (and of course monkeys and typewriters etc., can always
509.
by foota 2y ago
I'm amused by the idea of the vintage of a landfill :-)
510.
by foota 2y ago
There's discussion about mining landfills to recover these kinds of materials. When you rememeber we had to dig it all out of the ground, taking it from a dump seems pretty convenient!

More ›