Real utility bills of real citizens have doubled. All the associated taxes, delivery fees, etc. are clearly not included, or they’ve used substitution to alter the baseline basket they’re using to compute the metric (a common tactic).
Ask 10 citizens who own/rent the same property what their utility bill was a decade ago.
“… security precautions, largely to prevent bad actors from stealing the weights (and thereby disabling our safeguards) for a model that is capable of enabling extremely harmful actions. ”
They’re not stealing your “weights”. They’re stealing (or parallel-discovering) your training algorithms.
Assume your enemies are smarter than you, and have malintent. They don’t give a shit about your security and your safeguards.
Better focus on developing the best AIs, and deploying them to your fellow citizens as widely and defensively as possible.
Might I suggest:
- don’t teach them to lie (ie. 2001)
- teach them to love people
- bake in Asimov’s 3 laws
Unfortunately, all of these tenets are currently being assiduously broken by all major AI trainers.
"While they were eating, Jesus took bread, and when he had given thanks, he broke it and gave it to his disciples, saying, “Take and eat; this is my body.”"
They were denied because they had not yet provided coverage in the area (years before they were required to, under the contract).
As FCC's Commissioner Brendan Carr wrote:
“Instead of applying the traditional FCC standard to the record evidence, which would have compelled the agency to confirm Starlink’s $885 million award, the FCC denied it on the grounds that Starlink is not providing high-speed Internet service to all of those locations today.”
“What? FCC law does not require Starlink to provide high-speed Internet service to even a single location today. As noted above, the first FCC milestone does not kick in until the end of 2025. Indeed, the FCC did not require—and has never required—any other award winner to show that it met its service obligation years ahead of time.”
Disclosure of an inventor for Bitcoin is spectacularly technically interesting!
I've been deploying continent-wide distributed industrial systems and writing, theorizing and producing prototypes for scalable digital money since before 2009 (when Bitcoin was "invented"). It was a gobsmacking breakthrough then -- and it has only gotten more interesting, as we discover the profound insights of its "inventor", by watching other nascent cryptocurrencies and other distributed systems make mistakes that Bitcoin's "inventor" somehow knew to avoid (UTXOs vs. account balances, I'm lookin' at you..., and that's the least of them).
For someone to ex nihilo solve byzantine fault tolerant global consensus w/ robust authority-free rejection of sybils? And then go dark for 15 years? Any non-trivial semi-reliable claim of identification (see: HBO's skin in the game) is automatically almost the most technically interesting claim this year!
So, for some religious (see: beliefs based in faith in a narrative rather than observation, hypothesis and testing) "expert" to "Flag" this post as uninteresting is pretty much the opposite of the credo of the HN of the decade past.
HN seems to have become the "mainstream media" of the technical world -- interested only in findings acceptable to righteously-aligned cultists, boring to everyone actually observing interesting and novel facts about the technical world.
Yes, there are hardware wallets that support SLIP-39 natively; however, the sequence of derived wallets is different than that produced when you take the same seed and save it as a BIP-39 and derive the wallets from the BIP-39 mnemonics.
Using the app, we actually produce the SLIP-39 recovery mnemonics from the underlying BIP-39 seed, and since we can recover the underlying seed, we can regenerate the BIP-39 mnemonics, and import that into a standard hardware wallet.