What if we implement them but hide them deep in the settings or as experimental feature inside the hidden developer menu, behind multiple warning messages and password prompts? Only the very determined developers and advanced users would be able to unlock them. Then it's safe enough?
Users will unfortunately click on absolutely anything that a trusted (deservedly or otherwise) source tells them to, and you won’t be able to reliable convince them otherwise with UX alone. This includes all “developers only”, “click 5 times” etc. UX interventions.
You have to decide whether the feature warrants the remaining risk after all mitigations, or at least exceeds other, simpler attack vectors.
I think in this case it does, but it’s not an easy decision and I can understand most opposing positions as well.
Isn't AI the new hot thing, why are the miners still going after Bitcoin, when they can probably just use the same infra for AI and make more money, stay profitable.
"don't work well for AI" is a hell of an understatement, the Application they are Specific to is literally just sha256(sha256(x)), what AI are you going to do with that?
GP probably didn't mean that hardware though, but rather the facility, electricity supply, cooling, etc.
The disappointing truth is that we simply don't know. Satoshi never explained it. For SHA-2 it can be used as a mitigation against length-extension attacks¹, and this seems like the most likely explanation, but it's just speculation.
> The publicly traded miners have been adapting by diversifying into AI and high-performance computing, which offer more predictable revenue than mining bitcoin at a loss. Marathon Digital, Cipher Mining, and others have been building out data center capacity alongside their mining operations.
Yes, it does but lacks depth, the problem here is they are diversifying not pivoting and by virtue of game theory for each miner they stand to win on others exit, as the mining difficulty goes down, but this is creating a loss-loss situation.
That "Bam!" thing just brought Josh Starmer to mind. Anyone remember his book with the illustrated ML stuff? I used to watch his YouTube channel too. I really dig these kinds of explainers; they make learning so much more fun.
I agree to some extent, the average coders/programmers will definitely have a hard time finding jobs, but maybe they will start their own businesses instead.
We think of SQL as a high level language, an abstraction. We need to think about Code with LLMs similar to that as I still find LLMs to not be able to be completely autonomous at present.
There will be the need to debug understand the underlying execution at a low level just like in SQL or any other high level language.
There will be a requirement for humans to find better ways than LLMs and sometimes we might find frugal and ingenious ways to do stuff for which LLMs had taken the average solutions found in their datasets.
One of the 3 I's of innovation is iteration, if we call iterating innovation as building knockoffs then we might as well dismiss all other successful products that were iterations.
I completely agree, LibreOffice feels really complete these days and for the most of my needs it works well. The only concern I have is when I send it to someone who might not be using LibreOffice but say MS 365, is it still going to display what I had built and the way I had intended.
The best way to do that is via pdf, at least if the recipient doesn't need to edit it. I still have a $20 license for office 2024 for those few edge cases :)
It was inevitable, someone had to do it and I don't understand the awful fact that the West still lives in a bubble where it thinks that they are the only source of innovation in this world. There's a lot of innovation going on in the rest of the world too.
