How did you manage to uncover the operation? Is there any things that tipped you off? Is it like, accounts only posting on a precise subset of subs, or how much a network of accounts reply to each other? Or too much linking/references to products, or perfect spelling/identical consistent typos?
I'm curious because it feels like it could be built into a tool to analyse - even if it does become a bit of an arms race.
It's also a pretty silly thing to say difficulty = tokens. We all know line counts don't tell you much, and it shows in their own example.
Even if you did have Math-like tokenisation, refactoring a thousand lines of "X=..." to "Y=..." isnt a difficult problem even though it would be at least a thousand tokens. And if you could come up with E=mc^2 in a thousand tokens, does not make the two tasks remotely comparable difficulty.
Is this sarcasm? Not that Trump's word means anything, but Trump has been against it since his first term. Having cancelled it temporarily in that first term, has said that he'll end H1B if he gets reelected, and that US shouldn't have the H1B program.
It's only since 2025 when Elon was in his good books and told Republicans to not vote for a bill that Trump woke up that day and decided he'd be pro-H1B.
What are the quotes? I don't understand, some of them are just a single word like "Art", "Autoimmune", "Cartoons"; or innocuous phrases with no context.
The first two talks are in the "Ethics, Society & Politics" category, and the third in the "Art & Beauty" category. Why would they need to be about computing?
It's a big organisation, and politics is wrapped up in what they do, along with the post-WWII Antifaschism culture in Germany.
Even if it weren't the case, I don't get why attack them for helping stand up for democracy, something in dire need of advocacy these days
No I don't think it's the bits. I would say it's the computation. Inference requires performing a lot of matmul, and with more tokens the number of computation operations increases exponentially - O(n^2) at least. So increasing your context/conversation will quickly degrade performance
I seriously doubt it's the throughput of memory during inference that's the bottleneck here.
What do you mean? Is mRNA not used to produce the enzyme that these comments mentioned?
I don't think they were saying mRNA is gene editing itself. Just commenting on a modified mRNA helping the process compared to normal mRNA.
Might be misunderstanding though so correct me if I am
Japan's WWII history is uniquely bad but they don't learn about it.
Can threaten authors with treason for negative books like he did in an EO recently. Change school curriculums. Then Maga can start revising history..
Was the 2025 recession from tarrifs? Nah it was Biden's inflation, or Ukraine aid. Actually.. didn't China impose tarrifs on US and US just reciprocated?
Thank goodness the alternative media is bound by the fairness doctrine and don't choose how to cover a story, or if at all, and not incentived to reinforce their audience's beliefs and extremify content for views.
At least they ensure they are well researched on the matters they talk about. Right?
"without a valid reason" is doing a lot of heavy lifting here. Not only would this idea be impractical and highly subjective, determining what a valid reason is, is the same problem as defining the Law in the first place.
Can you insult someone? Can you say something wrong that you thought was right ("the lion cage is locked") that someone is injured from? What is their duties in checking the info they get is correct? Is there a min wage or not? What value is it? Does it change on city or state? Can under-age people sign contracts? Can they vote?
I'm interested for this reason too, even listened to AI TTS books before, but the issue is that they are very monotonous. The tone almost never changes, nor the pacing, it's all delivered with almost no variation which makes listening dull and easy to lose focus
Thanks! Design decisions in terms of backend architecture or the code or frontend design? One big one would be to offload many of the functions from the server to the client. I think validating game validity would still be possible but allow optimistically performing actions without latency
Actual Beej? Wow I remember absolutely loving reading your networking guide. It taught me so much and really showed me the depths and breadths of what can be done in code, how the net works (pun unintended), it was a great experience for me as a kid. Thanks! <3
I'm sure there's a reason, but it seems like an unusual use of Monte Carlo - it's all deterministic and there is no opposing player making choices. Must have something to do with uncertainties in projected orbits or imperfect simulations maybe?