Maybe Iām missing something, but has anyone actually justified this sort of āreasoningā by LLMs? Like, is there actually anything meaningfully different going on? Because it doesnāt seem to be distinguishable from asking a regular LLM to generate 20 paragraphs of ai fanfic pretending to reason about the original question, and the final result seems about as useful.
Possibly Iām the last to hear about this one, but seeing as proton mail has come up here a few times before: the founder and ceo Andy Yen is apparently a Trump fan.
Great pick by @realDonaldTrump. 10 years ago, Republicans were the party of big business and Dems stood for the little guys, but today the tables have completely turned. People forget that the current antitrust actions against Big Tech were started under the first Trump admin.
(from the beginning of december, on the nomination of trump staffer Gail Slater to antitrust post at the doj)
Apparently, the OpenMandriva folk (the inheritors of the venerable mandrake/mandriva Linux distro) are now best buddies with Bryan Lunduke (right wing tech grifter and q-anon fan) are decrying the left wing bias of Linux projects with a hilarious āwokeOS shellā
Archive of openmandriva forum post: https://archive.is/2025.01.11-001057/https://forum.openmandriva.org/t/came-here-from-lunduke/5516/1
Lovely juxtaposition of āletās stick it to the gay fagsā and āweāre accepting of everyone and thereās no hate hereā. Seems like a classy community all round. Itās a little sad to see how mandrake ended up, but there you go.
WokeOS here: https://web.archive.org/web/20250110234818/https://lindev.ch/wokeos.cpp
Itās pretty tedious and unimaginative. No idea who lindev are.
(eta: wasnāt me who originally found this, but Iām never quite sure whether itās ok to include sources for this sort of thing given the subject. on the other hand, the op has it as public post thatās been boosted a bunch of times, so here it is: https://tech.lgbt/@GeopJr/113807022917800887)
A real ceo does everything. Delegation is for losers who canāt cope. Canāt move fast enough and break enough things if youāre constantly waiting for your lackeys to catch up.
If those numbers people were cleverer than the ceo, theyād be the ones in charge, and they arenāt. Checkmate. Do you even read Ayn Rand, bro?
Remember that actual physicists can fall into the same trap, and believe themselves to be very smart too. Plenty suffer an irresistible urge to fix every other field thatās doing it wrong.
As an alternative to the various xkcds on the subject, have an smbc instead.
If it were merely a search engine, it risks not being ai enough. We already have search engines, and no one is gonna invest in that old garbage. So instead, it finds something that you might want thatās been predigested for ease of ai consumption (Retrieval), dumps it into the context window alongside your original question (Augmentation) and then bullshits about it (Generation).
Think of it as exactly the same stuff that the LLM folk have already tried to sell you, trying to work around limitations of training and data availability by providing ācut and paste as a serviceā to generate ever more complex prompts for you, in the hopes that this time youāll pay more for it than it costs to run.
And, whilst Iām here, a post from someone who tried using copilot to help with software dev for a year.
I think my favourite bit was
Donāt use LLMs for autocomplete, use them for dialogues about the code.
Tried that. Itās worse than a rubber duck, which at least knows to stay silent when it doesnāt know what itās talking about.
https://infosec.exchange/@david_chisnall/113690087142854474
(and also https://en.m.wikipedia.org/wiki/Rubber_duck_debugging for those who havenāt come across it)
Interesting article about netflix. I hadnāt really thought about the scale of their shitty forgettable movie generation, but there are apparently hundreds and hundreds of these things with big names attached and no-one watches them and no-one has heard of them and apparently Netflix doesnāt care about this because they can pitch magic numbers to their shareholders and everyone is happy.
āWhat are these movies?ā the Hollywood producer asked me. āAre they successful movies? Are they not? They have famous people in them. They get put out by major studios. And yet because we donāt have any reliable numbers from the streamers, we actually donāt know how many people have watched them. So what are they? If no one knows about them, if no one saw them, are they just something that people who are in them can talk about in meetings to get other jobs? Are we all just trying to keep the ball rolling so weāre just getting paid and having jobs, but no oneās really watching any of this stuff? When does the bubble burst? No one has any fucking clue.ā
What a colossal waste of money, brains, time and talent. I can see who the market for stuff like sora is, now.
For VPNs, at least, I can offer some suggestions. If you wanted to securely access a specific box or network of yours, tailscale is pretty great and very painless to use. If you wanted to do stuff without various folk noticing then thatās a bit trickier but Iāve been happy using mullvadā¦ theyāre not the cheapest, though they have some splendid anonymous payment mechanisms (you can literally mail them a wad of banknotes with a magic code on a bit of paperā¦ you donāt even need to muck about with bitcoin).
In further bluesky news, the team have a bit of an elon moment and forget how public they made everything.
https://bsky.app/profile/miriambo.bsky.social/post/3ldq2c7lu6c25 (only readable if you are logged in to bluesky)
Blueskyās approach to using domain names to mean identity is now showing cracks that everyone can see: https://tedium.co/2024/12/17/bluesky-impersonation-risks/
(it was always shaky, but mostly only shown by infosec folks who signed up as amazon s3, etc)
TL;DR: scammer buys .com domain for journalistās name, registers it on bluesky, demands money to hand it over or face reputational damage, uses other fake accounts with plausible names and backgrounds to encourage the mark to pay up. Fun stuff. The best bit is when the sockpuppets got one of the real people they were pretending to be banned from bluesky.
Nvidia doing their part to help consumers associate AI with unwanted useless bloatware thatās foisted upon them.
I will find someone who I consider better than me in relevant ways,
Lemme guess, rich, white, asshole? (now I write this, I realise it could be about the author of the blog post too, and not just the bull heās seeking).
These people continue to be so utterly delusional about the nature of success. The desperate need to believe that genetics is destiny, and that the ultra-wealthy got that way because they are also ultra-competent instead of merely being ultra-lucky and/or ultra-rapacious.
I guess the future is a race to see what comes firstā¦ the ultra-wealthy habsburging themselves into oblivion, the oceans boiling, or a resurgence in the construction of hand-built artisanal tumbrels.
Funny how right-wing types reaaaaaly love unborn kids. Theyāre the best focus group for almost any policy!
Whatās a hard quality of life indicator?
Interesting post and corresponding mastodon thread on the non-decentralised-ness of bluesky by cwebber.
https://dustycloud.org/blog/how-decentralized-is-bluesky/
https://social.coop/@cwebber/113527462572885698
The author is keen about this particular āvision statementā:
Preparing for the organization as a future adversary.
The assumption being, stuff gets enshittified and how might you guard your product against the future stupid and awful whims of management and investors?
Of course, they donāt consider that it cuts both ways, and Jack Dorseyās personal grumbles about Twitter. The risk from his point of view was the company he founded doing evil unthinkable things like, uh, banning nazis. Heās keen for that sort of thing to never happen again on his platforms.
Itās a long read, but a good one (though not a nice one).
Someone on mastodon (canāt remember who right now) joked that they were expecting the c++ committee to publicly support trump, in the hopes he would retract the usg memory safety requirements. I can now believe that they might have considered that, and are probably hoping heāll come down in their favour now that heās coming in.
Just take the same approach that mathstodon does with their latex formatting, to wit: just barf the probably-human-unreadable format strings out over activity pub, and shrug when people complain because what else are you gonna do?
It isnāt like inline images work any better over activitypub in general (and even markdown/html is tricky, in the face of various mediocre mastodon client apps, some of which I use) and doing server-side conversion of latex (or whatever)to mathml is a nonstarter because of it will just be filtered out by clients anyway as unsupported and possibly malicious.
The least bad option might be something like asciimath (which can be prettyprinted via mathjax client javascript in the web view of the site) as its raw form is less awful to read than latex, and so would be more suited to viewing via activitypub or in a lemmy app.
I wonder what proportion of awful.systems users view it via the web interface.
Everyone and their dog uses mathjax or katex to render math client-side these days. Iām not gonna say that it would be trivial, but it also shouldnāt require you to get elbow-deep in lemmyās own post-formatting code.
Unfortunately, thereās no obvious prior art that can be stolen here, and thereās only one slightly confused feature request in the lemmy-ui repo, so someone is going to have to nail all the bits together themselves.
Sounds like all it would take is one company to do it right, and theyād clean up. Except somehow, with all of the billions being poured into it, every product with ai sprinkled on it is worse than the non-ai-sprinkled alternatives.
Now, maybe this is finally the sign that everyone will accept that The Market is completely fucking stupid and useless, and that literally every company involved in ai is holding it wrong.
Or, and I know itās a bit of a stretch here, but consider the possibility that ai just isnāt very useful except for fooling humans and maybe you can fool people into paying for it but itās a lot harder to fool them into thinking it makes stuff better.