whereismyacc

5 ms·

1.
by whereismyacc 8mo ago
That I can absolutely believe but the big competition is in enterprise gpt-5-size models.
2.
by whereismyacc 8mo ago
best inference silicon in the world generally or specialized to smaller models/edge?
3.
by whereismyacc 11mo ago
What would be the evolutionary advantage of this? I know nothing about biology so I'm struggling to see the use of such a network. Some kind of synchronization?
4.
by whereismyacc 1y ago
yes but there could conceivably be better tricks to discover
5.
by whereismyacc 1y ago
> There are no probability distributions over possible states when there is perfect knowledge of the state. I know very little about physics but I thought that the leading interpretations of quantum physics say that the probability distr
6.
by whereismyacc 1y ago
It sounds like you're talking about information entropy which to my understanding is analogue to but not the same as entropy in physics?
7.
by whereismyacc 1y ago
You were still going to lose export markets as international competition grew, and you were still going to shift to higher value-added service jobs as the economy developed.
8.
by whereismyacc 1y ago
America got richer and outgrew the phase where tons of factory jobs made sense. It seems pretty clear to me that well-paying manufacturing job in developed countries were the product of a particular moment in time where poorer countries cou
9.
by whereismyacc 1y ago
The question was about what happens in other cases.
10.
by whereismyacc 2y ago
to understand what they said, or to understand a proof of why it would be true? any stats class would be enough to understand what they said
11.
by whereismyacc 2y ago
welp i guess i should get my sight checked
12.
by whereismyacc 2y ago
i thought it was really cool when it picked up the grapes by the vine edit: it didn't.
13.
by whereismyacc 2y ago
Oh yeah, I don't think I would suggest actually implementing this because it's just a lot of costs and fees to emulate what otherwise would be an index without the stock
14.
by whereismyacc 2y ago
Yeah of course you can't escape the general market risk, i'm only talking about a strategy to avoid net exposure to risk specific to the TSLA stock.
15.
by whereismyacc 2y ago
I definitely don't know much about finance, but a short is essentially being obliged to sell a position at a future date, right? You match your short position against the fraction of your ETF which is in TSLA, such that you are obliged
16.
by whereismyacc 2y ago
well not if the short matches the amount you hold in tesla stock through the ETF right?
17.
by whereismyacc 2y ago
I really don't like that these models can be branded as Deepseek R1.
18.
by whereismyacc 2y ago
I've never heard about Kagi but a paid search engine just sounds great? I assume the userbase will always be small enough that websites won't bother doing SEO for it. Maybe there's some low hanging fruit there in getting less
19.
by whereismyacc 2y ago
sounds good
20.
by whereismyacc 2y ago
They literally pay the tax but yes often the incidence ends up on the consumers.
21.
by whereismyacc 2y ago
Clearly it could never be enough to draw that conclusion but it might be very weak evidence in one direction
22.
by whereismyacc 2y ago
I've gotten the impression that: 1. The bias is mostly due to the training data being from larger models, which were heavily RLHF'd. It identified that OpenAI/Qwen models tended to refuse to answer certain queries, and imitat
23.
by whereismyacc 2y ago
It could be that, but also many other languages don't capitalize titles the way English does. It took me a long time as an English learner to learn that it was expected. It still feels very weird to me.
24.
by whereismyacc 2y ago
Aha, for some reason I thought they provided full-size Llama through some bundling of multiple chips. Fair enough then, anyway long term I feel like providers running powerful open models on purpose built inference ASICs will be really awes
25.
by whereismyacc 2y ago
Neither of the deepseek models are on Groq yet, but when/if they are, that combination makes so much sense. A high quality open reasoning model, but you compensate for the slow inference of reasoning models with fast ASICs.
26.
by whereismyacc 2y ago
At least (I assume?) a nice feature of short-form video app algorithms is that every few seconds, the users swipe or like or whatever to reveal preferences. The rate at which that data comes in from users is much higher than for long-form p
27.
by whereismyacc 2y ago
The weights are quantized down to fewer bits in order to save on memory. The quantization loss is going to result in worse generations.
28.
by whereismyacc 2y ago
I feel like this gets at the core issue which is food being too palatable and not satiating enough? At least on the obesity / calorie overconsumption front, I think it's just that simple. 'ultra-processed foods' or &#x27
29.
by whereismyacc 2y ago
I figured like thirty seconds into using the site that resizing it smaller would give me more DVD bounces per second. But then during resizing i kinda cheated myself some points accidentally, and discovered that trick where they're jus
30.
by whereismyacc 2y ago
I'm just responding to the portion where you said the only one that benefits is amazon. Surely at the very least, the immigrants themselves benefit if the alternative was no ability to work in the US?

More ›