svnt

28 ms·

451.
by svnt 2y ago
The marketplace for this, as compared with generic software developers, would be so small and infrequently used as to be basically nonfunctional. I would recommend searching for terms of art such as vhdl and verilog in normal freelancer sit
452.
by svnt 2y ago
My argument wasn’t that the well-funded entities were optimizing to reduce training costs, but the opposite: they need creative ways to spend $1B that provide some tangible advantage. But they need operating costs to be low or they lose mon
453.
by svnt 2y ago
There are at least two different optimizations happening: 1) optimizing the model training 2) optimizing the model operation The $1B-spend holy grail is that it costs a lot of money to train, and almost nothing to operate, a proprietary mod
454.
by svnt 2y ago
Is it still even worth the electricity to do this on a GPU? It wouldn’t surprise me if some startups were renting them out, but is anyone still mining any volume of crypto on GPUs? edit: I guess to your point if it is not knowingly then the
455.
by svnt 2y ago
No, you need two of the newest M3 Macbook Pros with maxed RAM, which in practice some people might have, but it is not gettable by using old hardware. And not having tried it, I’m guessing it will probably run at 1-2 tokens per second or le
456.
by svnt 2y ago
I’ll flip this around a bit: If I’ve raised $1B to buy GPUs and train a “bigger model”, a major part of my competitive advantage is having $1B to spend on sufficient GPUs to train a bigger model. If, after having raised that money it become
457.
by svnt 2y ago
15g of explosives is sufficient, probably, based on the analysis I’ve seen. 15g of explosives is not very big. The speaker is another potential home, that is a good point. Big cone and heavy magnetics. This display weighs 15g, and uses chea
458.
by svnt 2y ago
Nothing, they aren’t looking for 2”x1” sheets of copper within electronic devices, and presumably the thin layer of explosives would be sealed and washed.
459.
by svnt 2y ago
The viability and profitability of the strategy depend critically on the ownership and control position of the CEO. Many CEOs do not have all that much stock, in percentage terms. If they have access to resources and sufficient board contro
460.
by svnt 2y ago
The best evidence we have now suggests that the devices used had removable (AAA) batteries, not built-in batteries. If I was buying pagers and had previously been hit by intelligence ops I would be buying batteries in random supermarkets.
461.
by svnt 2y ago
My best guess is explosively formed penetrator in the display. I don’t think wholesale replacement of the pagers was likely to work for a number of reasons. They had to go one step up the supply chain. The EFP display could be set to trigge
462.
by svnt 2y ago
You probably need firmware and some major component modification, such as a display or battery, but not more than this, to pull it off. So at a minimum, two components, or perhaps one smart component such as a display. It seems the model wa
463.
by svnt 2y ago
I don’t know if that is true, unless your definition of qualia is so loose that it includes (or emerges from) basic data representations such as vectors comprised of weights and vertices, in which case I think LLMs could perhaps be argued t
464.
by svnt 2y ago
I think it is like qualia in the sense that it can be something known experientially, but spatial and logical reasoning are things that transcend subjectivity. No doubt LLMs do not have qualia as humans do, but I also think that the diversi
465.
by svnt 2y ago
Based on my understanding from the description, it is ~8x faster (250 GFLOPS) for vector ops (vs. SVE mode at 31 GFLOPS which is CPU-ish) and 60-100 times faster (e.g. 2005 GFLOPS) for matrix multiplication for single-precision values.
466.
by svnt 2y ago
An example is where the link between the ideas is not present in the prompt. E.g. an oversimple example is: It’s going to rain tomorrow so you should take an umbrella. If someone doesn’t understand the relationship between an umbrella and k
467.
by svnt 2y ago
People really love it when you tell them what breed their dog is right after you’ve looked it up on your phone.
468.
by svnt 2y ago
This is a joke, I think. It is a joke, right?
469.
by svnt 2y ago
The idea that it is always better to invest in potential efficiency improvements in solar vs just installing more of what exists.
470.
by svnt 2y ago
I had this theory that generations raised on the internet and exposed to it from birth would be the most humble generations ever, because we all look for ways to be uniquely valuable, and it became nearly impossible to be egotistical when f
471.
by svnt 2y ago
> Assuming all else is equal, it's literally free energy. When comparing silicon and perovskite (and probably any other material) this is a bad assumption. Since this is a bad assumption, the rest of your position falls apart.
472.
by svnt 2y ago
When github started, for most people the only purpose was just so you didn’t have to manage a server holding your repository. To avoid using it at that point for private projects required all of ssh and a $5/mo virtual machine somewher
473.
by svnt 2y ago
To get a better picture of how this comes about: Microsoft has a study for their own internal product use, and wants to show its efficacy. The results aren’t as broadly successful as one would hope. Accenture is the kind of company that co
474.
by svnt 2y ago
This kind of conformist thinking leads to bad outcomes.
475.
by svnt 2y ago
I worked there on a contract while they were building a phone and had as many as six managers over six months. One of them would come in at 11, talk to me about nothing until lunch, then come back after lunch, talk more, and go home. I had
476.
by svnt 2y ago
Transformers in 2017 as the basis, but then the quantization-emergence link as a grad student project using spare time on ridiculously large A100 clusters in 2021/2022 is what finally brought about this present moment. I feel it is fai
477.
by svnt 2y ago
> Risk and reward are intrinsically linked There are innumerable ways to increase your risk without increasing your potential reward.
478.
by svnt 2y ago
Presumably you work in the field in North America, and not South India or Africa.
479.
by svnt 2y ago
It can also cause the opposite harm: a requirement for treatment of something that isn’t actually broken, just because it has been culturally declared to be broken.
480.
by svnt 2y ago
Hastings is rather unique in the current founder era in that he had previously founded a company, recognized he did not have the ability to manage it, attempted to step down, and then was told by his board to figure it out. He went on to es

More ›