696 karma · joined October 16, 2016
distributed hash tables, p2p systems, supply chain and network security.
I also deeply hate NPM. (Normalizing NPM is a hazard to society) So maybe I'm biased...
"Would it explode or just punch a hole" depends on the energies of the impact involved. If there is only enough energy to break chemical bonds, it would punch a hole.
As energy of the impact gets higher the hole gets deeper. Probably punching all the way through the ship.
Once the impact energy gets strong enough to break nuclear bonds, the story changes. Now we have an e=mc^2 event at the impact site. It isn't just the impact energy anymore. At minimum, the punched hole just got wider and potentially radioactive as byproducts go on adventures. Once the hole is wider than the ship, I think that counts as "exploded".
Ask any therapist, consciousness has limits, it has bound depth. It takes a lot of work to re-train it. It isn't a magical human only special thing.
We discovered half the planets by doing the math to predict the orbits based on the known distribution of math in the solar system.
General Relativity was initially validated by predicting mercury's orbit accurately.
People aren't realizing just how financially tight and desperate poorer people are getting right now
3) make internal state not useful to the attacker.
4) assuming the ciphertexts won't leak seems silly. Might as well hand them out.
Which leads to what they call "trial decryption" to be a better solution. If you are that worried about the scalability of your whistleblower protocol at that level, we are trying to solve the wrong problem.
Literally scraping the largest accessible corpus of human communication and attempting to emulate it and placing it as a MITM attack between us.
Also, the AI has a finite size, which means it has to be scaled up for bigger graphs. I doubt it would work at all for non-euclidian 2d graphs.
The exciting thing is that they made an AI do a toy problem, not that it was efficient.
But people are dumb and tech companies have brain control over them. "AI good and solve all problems"
It reminds me of the "slime mold solves mazes" thing. It was interesting and there were even things to learn from it, but it wasn't magic and couldn't give us better general tools.
In the US, basically all of Central American is coming to live here in the coming decades. If we don't find a way to invite and welcome them, our society is going to break down.
Conservatives are freaking out about "invasions over the southern border" and advocating killing people. This is just the start. This is the trickle. No amount of "border security" will stop people fleeing for their lives en-mass. The US just isn't culturally prepared for this and I don't know how we can change it.
What's the specificity?
The biggest thing you seem to miss, is that being a sociopath is a lack of intuition and instincts, not what choices they make. Our entire society and economic structure is about convincing sociopaths to contribute despite the disability.
Getting himself fired was something he knew would be part of this protest.
I'm really interested in this concept but the above link is vaporware.
None of these projects actually have a noun phrase I can search that unify them. I don't even know what to call this flavor of kinesthetic computing.
Basically you can barely trust publicly scrutinized science to not mess up data gathering. I definitely don't trust companies that sell automated tools for user satisfaction to actually have good metrics for it.
You see it is growing quickly. You could productionize now, or invest a bit longer and make it 2x better.
An independent problem is that your model is really powerful, but left to it's own devices it spouts the worst of 4chan right alongside high quality reddit comments.
When do you ship? You have 2 criteria: After you yoke it to not be a Nazi and once returns on quality investment start to fall.
Both Google and OpenAi hit the point where returns started to diminish. It turns out the hard part was yokeing it. OpenAi, being more agile, throws massive pools of borderline slave labor at the problem while Google is still ramping up their "let's just throw supervised training at the problem" solution.
These products were released past the inflection point of the sigmoid. The marketing is strong to re-create the old moore-law hype to re-loosen investors wallets, but the return quality investment is already diminishing. The primary way it is changing is in finding composition of tools and new use cases.
Clean the marketing scales from your eyes and feel welcome to the plateau.