Discussion
just rereading this and basically Amy and I completely explained this AI shit the first time in 2023
https://davidgerard.co.uk/blockchain/2023/06/03/crypto-collapse-get-in-loser-were-pivoting-to-ai/
everything since is because nobody fucking listened
so much for sweet reason and debate
@davidgerard It's crazy this went so far. Super intelligence curing cancer and climate change, if we just built out enough data center capacity to starve people into the extinction. News at 11.
@davidgerard Chomsky made his view known in exactly the same time frame, when asked.
https://chomsky.info/20230503-2/
@davidgerard I'm seeing worrying signs in the internet chatter this month that they're trying to decide whether to pivot [back] to quantum computing or go straight for commercial fusion reactors. I suspect the latter is currently winning because "we can build cheap small fusion reactors to power the AI data centres" is an obvious follow-on spiel, while nobody quite understands the gimme with quantum.
@cstross @davidgerard I thought the grift was growing clusters of human brain cells and implying that they’re being put to work as post-AI agents, with the man behind the curtain being some combination of LLM slopbots, old-fashioned algorithms and third-world sweatshops.
@acb @davidgerard There is no one grift any more, just a thundering herd of sociopathic con men (and some women: eg. Elizabeth Holmes) trying to figure out which is the next bandwagon to jump on.
@cstross humanoid robots!
@davidgerard Definitely a novelty item! Not terribly practical but they'll be filling in for care assistants in nursing homes and hospitals *any day now*. Once they stop disintegrating and catching fire when they run into walls, that is.
@cstross
Ahhh, that explains the Sabine Hoessenfelder piece on why quantum doesn't work...
(There's must be a bit of "not yet bro, just wait a bit, bro" going on)
(Yes, I think she's a suspicious source on a number of topics)
@davidgerard
The pivot to commercial fusion power plants is already mid-swing. See the plenty of YT videos, and even in VC-averse Germany start-ups like Proxima Fusion popping up.
Proxima Fusion is a spin-off from the MPI for Plasma Physics (i.e. they have the Wendelstein 7-X knowledge).
I toured that thing a year ago. The "auxillary" tech carries a lot more complexity than the actual plasma confinement vessel. Things like plasma heating are non-trivial.
@cstross I had hopes last month when the first Trump-and-dump SMR schemes went bust that this branch of the tech tree of scams was already folding.
@henryk NUCLEAR WILL NEVER DIE!!!!!11!!!1!!ELEVENTY!!!
Seriously, nuclear projects are inherently large, which means expensive, which means they're a money vortex, which attracts a certain kind of personality we've been breeding for since the Thatcher/Reagan revolution and now we've raised a ripe crop of shit-headed grifters.
@davidgerard @cstross I'm not even 100% sure on that. There's a short-term and a long-term argument. On the short term, nuclear (neither conventional nor SMR) will not be ready to affect anything. On the long term, *any* form of fusion -or- fission reaction adds heat to the earth system. And has the fuel problem.
So, if you're building small: Why bother, solar would have you covered in the same time five times over.
If you're building big: you're warming the planet and need fuel from somewhere.
@davidgerard @henryk The main ongoing use case for nuclear that I can see is for decarbonizing shipping. The alternative zero carbon power sources are all a bit crap (synthetic fuels: energetically inefficient, ammonia: poisonous as fuck, batteries: range-limited, sails: don't scale up to ULCC size, don't work if no wind).
I assume we will DESPERATELY need bulk shipping this century as our agricultural stability craters, to ship food from still-productive regions into famine zones.
@cstross @davidgerard @henryk I spent five years researching SMRs and nuclear power and the grid in general as a hobby. My conclusion is tech billionaires are funding nuclear reactor startups 90% as a counter argument to data center electricity use by their main businesses, and 10% because of the "this is cool as shit" factor. (SMRs are cool as shit, and also completely impractical and dangerous.)
@vaurora @cstross @davidgerard @henryk
My (largely unresearched) concerns about SMRs is you can make the core small and modular, but all the shielding, containment, security, steam generation and cooling benefits from scale and is big for good reasons.
Fusion powered Quantum AI, when IPO?
@davidgerard @expertenkommision_cyberunfall @cstross Said in the same tone of voice as Sark from Tron saying "on the game grid"
@cstross @davidgerard "quantum AI" just ticks so many bullshit boxes that it'll be really hard to avoid latching on to...
@jon_ellis @davidgerard Some genius scam artist is going to latch onto Roger Penrose's crazycakes quantum brain hypothesis from "The Emperor's New Mind" as the grift for their new AI project. Any day now.
@cstross @davidgerard have to admit that by 'shadows of the mind' i'd got to the point of admiring the hustle, but at the same time you have to admit that at least he was proposing a mechanism and wasn't just falling back on "and at scale consciousness emerged". that feels like the real woo woo to me!
@cstross @jon_ellis @davidgerard
Oh geez Penrose. God damn it man I had just about recycled those neurons and you refreshed them again.
@davidgerard has anyone put an llm in a smart contract yet?
@parsingphase they have started vibe coding the smart contracts, which is just what you want to do with something you're pretending is money
@davidgerard
AFAIK, smart contracts didn’t work very well before LLMs. So that might be a smart move.
@davidgerard @parsingphase It's worth pointing out that Bitcoin Core is using a certain RPC library, but not leveraging its security model properties in any way, just knocking stuff through it in the clear.
@davidgerard June 2023 feels like 250 years ago.
@davidgerard The past three years have been the longest two decades of my life.
@davidgerard crypto needed mostly speculation to invent a purpose. AI already has mundane ones: translation, free accessibility, helping non coders to explore ideas, research, education, automation.
If used right its a solid hammer. But humans like to use it as a casino. Like houses, food, water, sneakers, yeah literally everything. The interesting question isn’t whether the hammer is evil. It’s who profits from these fairy tales. Blaming the object is a waste of time.
@davidgerard quite the opposite. I’m saying the social context is the interesting part.
A technology being developed under capitalism doesn’t make its current ownership structure, business model or use cases intrinsic properties of the technology. Otherwise you could make the same argument about the internet, pharmaceuticals, agriculture or housing.
Declaring the whole AI category rotten actually removes social context rather than adding it.
@ButchCoolidge@mastodon.social except in practice. you are not bright enough to kramer into my threads.
@davidgerard "AI is a solid hammer" no idiot it's a hammer made up of oil goop... Damn, the confidence of these buffoons when repeating the same stupid shit the millionth time as if it was the newest sharpest argument in the Intellectual Battleground Of Big Minds...
@ButchCoolidge this assumes a technology without a social context is possible, and decontextualising "AI" from its circumstances and reasons to exist is a standard and odious form of slop apologetics
@davidgerard @ButchCoolidge “AI cannot be separated from its social context.”
insults disagreement, blocks dissent.
Well. There’s your social context. And the reason why AI has no moral.
Maybe the machine isn’t the only thing worth examining
@davidgerard It's depressing how well this aged, especially since it was all so easy to predict. Que the pivot to "quantum computing" in 3... 2...
@tiredgothgirl we barely knew anything before we researched that and it wasn't hard to get it all correct!
the failing is not knowledge
Install bonfire.cafe
Get the full app experience
