Yes but you have a) no idea how bad the suck will be b) no idea how long the suck will last and c) what the "better" will look like.
I am not advocating that the status quo should go on, I am just saying that the repercussions will be vast and long lasting, likely resulting in a world war, and whatever follows after that(if anything).
> To be honest, I still don’t fully understand how ideology can cloud the mind so thoroughly that only a single way of thinking remains possible.
From what I know, and please correct me if I am wrong; it relates to fear and cognitive dissonance. First, by creating FUD the perpetrator can cause physical narrow-mindedness within the brain, the amygdala — centre of emotions if you will — takes control which reduces reasoning capabilities. Second, by introducing multiple conflicting viewpoints in that state, you induce what we call cognitive dissonance. The brain is unable to reconcile the two opposing (or even just differing) views. This is a conflict at the circuit level of the brain, and the brain needs to reach a conclusion, and conveniently the conclusion is produced by the perpetrators of fud, those who seek to control/exploit others.
China is on the path to doing just that, and I expect the heads of the government (or at least the people behind the scenes) in both EU and India consider this.
The US has, unfortunately, proven to be a very unreliable partner.
The problem isn’t only the U.S. China is also unreliable, and Russia as well. Just look at what happened to the EU with its dependence on Russian gas: once the war in Ukraine started, that dependency immediately turned into a major vulnerability.
There are plenty of GPT evangelists too on HN. Plenty of people who have a few successful interactions with a code model and extrapolate that to LLMs reasoning and solving general problems etc.
Any reason for creating a new tensor when accumulating grads over updating the existing one?
Edit: I asked this before I read the design decisions. Reasoning is, as far as I understand, that for simplificity no in-place operations hence accumulating it done on a new tensor.
yeah, exactly. it's for explicit ownership transfer. you always own what you receive, sum it, release both inputs, done. no mutation tracking, no aliasing concerns.
That's the (obvious, I guess) comparison I was thinking of too, but IIRC correctly the issues there were 1) allegations of bribes (which ended up being false/that witness arrested by the FBI for lying about it) and 2) Biden improperly leveraging the State Dept (which was also found to be untrue by two different Republican Senate investigations).
Now if the issue was Hunter Biden being on the board at all -- even if independent of any Joe Biden dealmaking -- then I'm very curious how the Republicans sounding alarms back then react to the Barron Trump TikTok board seat now.
did I miss the news that the US government forced a deal that transferred partial ownership of a foreign company to Burisma, and put Hunter on the board?
The logic we typically use for repeaters (EDFA, erbium-doped fiber amplifiers) for long-distance lines amplifies but does not clean noise (so across the oceans, you are very much bound by SNR). And you need one of them every 80 km or so in typical fiber.
