this isn’t surprising, but now it’s confirmed: in addition to the environmental damage generative AI does by operating, and in spite of all attempts to greenwash it and present it as somehow a solution to climate change, of course Microsoft’s been pushing very hard for the oil and gas industry to use generative AI to maximize resource exploitation and production (via Timnit Gebru)
TechTakes
Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.
This is not debate club. Unless it’s amusing debate.
For actually-good tech, you want our NotAwfulTech community
tbh i don't see a single sane way that genai could be used for anything like they say it can be, if it works it's gotta be something more or less custom. but ms doesn't care, because they're selling shovels so it doesn't matter if their shit doesn't work as long as someone's buying. it sorta starts looking like cryptobros in 2020-ish trying to insert themselves as middlemen everywhere where there's already some money
remember all the fucking rubes saying Proton’s LLM wasn’t a problem cause only business and visionary accounts had access to it? well, only one month later of fucking course they went back on that and now it’s included with duo and family accounts, and my soon to be cancelled unlimited account just popped an ad for it on the compose window trying to get me to opt into the free trial for the fucking thing (and also the button’s purple just as a last dark pattern to try and fool users into clicking it)
Why do I get the feeling we're gonna see a colossal tech crash
@self I wonder what "popular demand" and "overwhelming number of requests". Are there actually a lot of people asking for this? Are there more than there are people begging them not to? Maybe I just live in an anti-AI bubble, because I sure don't encounter a lot of pro-AI views.
it’s not just you — I can dig up the posts if you’re curious, but Proton wrote their last user survey so it was impossible to say no to this LLM crap directly, and they still got caught massively fudging the numbers to make this needless bullshit look popular. I can promise you it’s just the same people doing that again, except this time there’s no publicly accessible numbers they can be called out over
@self I can understand Microsoft and Google chasing this fad, but I would expect most of Proton's users to be exactly the type of person who's against this. But of course it's easy to fudge with numbers.
I would expect most of Proton’s users to be exactly the type of person who’s against this
that’s very true! unfortunately, we’ve discovered that a lot of the foundational members of Proton’s board and engineering team are huge LLM fans (and gigantic Bitcoin fans too — that’s why Proton released a Bitcoin wallet, of all things, almost simultaneously with this LLM bullshit)
we’re not sure if something changed that suddenly made them go all in on their bad ideas, but the initial communication around Scribe was how much Proton’s business users wanted it — and the survey was very much crafted to get what looked like a pro-LLM response from that demographic. Proton has essentially admitted that they’re doing this for their tiny number of enterprise whales rather than their normal privacy-conscious users; it’s a shame they’re willing to burn their business down for that kind of short-term gain. I can only imagine them enabling the LLM for all their paid accounts this quickly is either a desperation move because the feature didn’t do the numbers they hoped for, or it’s a sign that Proton’s otherwise compromised.
is this the bad place?
Yes.
Ffs I just swapped to Proton for drive and email. Thankfully only done a couple of email migrations.
Its worse for me - I've got a metric shitload of emails on Proton. Thankfully, I'm not using them for anything particularly important.
from the last time this came up, Tuta is of the few that aren't, although there's not really anyone with feature match on some of proton's features afaik
I agree; Tuta is the only real replacement, and they’ve promised (for what that’s worth) they don’t have any plans for AI features. I may migrate to Tuta myself, but I can’t truly recommend it — as always, I have to point out that Tuta is still a single point of failure like Proton, and one day I hope we’re able to design a federated, e2e encrypted replacement for email (that crucially isn’t gpg or anything like it — imagine teaching your grandma and your drug dealer (assuming they’re not the same person) to use that kind of thing)
@froztbyte
I forgot I'm using their VPN too, and they accept bitcoin (not a big deal, but useful). I was aware of Tuta and fastmail when I chose Proton.
For now I'll prioritize moving to a self managed domain to make swapping provider easier in future.
Apparently it implements chain-of-thought, which either means they changed the RHFL dataset to force it to explain its 'reasoning' when answering or to do self questioning loops, or that it reprompts itsefl multiple times behind the scenes according to some heuristic until it synthesize a best result, it's not really clear.
Can't wait to waste five pools of drinkable water to be told to use C# features that don't exist, but at least it got like 25.2452323760909304593095% better at solving math olympiads as long as you allow it a few tens of tries for each question.
Would there ever be a way to tell that they didn't just feed the answers into the training data?
Habsburgs are good, actually: https://news.ycombinator.com/item?id=41497419
Maybe it's because I've started reading a book about Germany and Austro-Hungary in WW1 (Ring of Steel) but I've suddenly started pattern-matching a bunch of pro A-H comments in HN. "It was a peaceful multi-national nation" well yeah until they pointlessly insisted on invading Serbia (and fucking that up twice before being bailed out by Germany) thus setting of the wider war. And when refugees from Galicia had to flee the Russians they were not happily accepted by the rest of the Empire.
Anyway, A-H was teetering on the edge before WW1 and signed their own death warrant willingly.
As always in HN you can find links to new horrifying examples of fascism: https://theworthyhouse.com/2021/06/17/the-foundationalist-manifesto-the-politics-of-future-past/
We need a high-quality Guards! Guards! movie so that monarchists can be given the Clockwork Orange treatment with it.
oh boy, some of the comments there
I think that the people in the breadline would have better things to do than be rude on twitter. Or maybe not, it is a fun way to pass the time for some.
how to instantly identify someone with a social circle barely stretching past their own nose. and I’d bet there’s some “work = moral” thinking held there, too
When [musk’s new] supercomputer gets to full capacity, the local utility says it’s going to need a million gallons of water per day and 150 megawatts of electricity — enough to power 100,000 homes per year.
The locals might get ornery about this, and I'd be willing to make the trip to help.
There are some steam turbine power plants (like coal-fired) (on smaller side) with power output like that