Stock market chaos as China reveals secret snake oil refinery, threatening US snake oil monopoly.

https://mstdn.social/@Nickiquote/113901910776808096

Snake oil yesterday, magic beans today.

AI to hit 40% of jobs and worsen inequality, IMF says

https://www.bbc.com/news/business-67977967

Going to hit 40% of jobs in 5 years the next

Is it magic beans and snake oil, eg. a scam or is it a legit threat, you can’t have both

we’re not in the slower than walking era, we’re in the ford model t era, i’m reading someone call the ford model t magic beans (eg. a scam) while millions of people are driving one every day and simultaneously worried it’s going to ruin transportation all over the north america

“It took years of hard work for us to steal beans from farmers, apply our unique brand of magic, and seek investment from our nation’s finest rubes and oafs,” a Beanco spokesman said. “Now DeepBean wants to steal our magic beans, rebrand the magic, and get money from their own buffoons and clods? It’s just not right.”

again, these are contradictory statements, it cannot be both a scam that has no value (eg. magic beans) and going to take all our jobs

But cars have always had value, we’re talking under an article that calls AI “magic beans”

How could magic beans which produce garbage enslave humanity under capitalism? Would you argue that cars which replaced horses as the primary mode of transport in a few years be called this?

You can argue that AI does some things badly, it’s still very very early on and the progress people are making is insane, like nothing I’ve seen before, but you can’t argue it is worthless and a giant threat to us at the same time, this is contradictory

That can’t be possible, if they fire all their workers and produce AI slop then we will simply start a new company with human workers that doesn’t.

This isn’t communism we can simply start a new business at any time for any reason, with or without AI

companies are getting ahead of themselves about the opportunity to fire human employees

But then if they produce garbage with AI people will buy the non-garbage product

Either it produces something of value or it doesn’t, if it’s producing garbage, lowering output, etc then it’s not a threat to our jobs because most people don’t like garbage, if it’s producing genuine value then it will be.

i have noticed that there are two competing narratives in the leftwingosphere:

A) ai is 100% slop garbage and a giant waste of electricity, pumping out garbage images with multiple hands and the text is nothing but hallucinations that can’t even count the number of r’s in “strawberry”

and at the same time

B) AI is going to take all our jobs and we will all be homeless and poor while tech billionaire CEOs turn us into slaves

She asked, “Why do men feel so threatened by women and others who are finally getting a seat at the table?”

She admits the competition has increased (and with women making up 50% of the population competition has damn near doubled), and is wondering why men are getting more competitive?

That script says all we need to do is provide for and protect our family. **Yet, the majority of women today also work and provide. **

Which has never been the case in history before, who knew if you give women the same rights and freedoms as men (and added bonuses like DIE to promote them over men) things would break down??

yeah sounds like bs to me

but we had the same thing with Alpaca, Llama2, Llama3, 3.2, Mistral, Phi…

I don’t believe so, or at least, them all getting smaller and/or more intelligent isn’t the point, it’s how they did it

I noted above that if DeepSeek had access to H100s they probably would have used a larger cluster to train their model, simply because that would have been the easier option; the fact they didn’t, and were bandwidth constrained, drove a lot of their decisions in terms of both model architecture and their training infrastructure. Just look at the U.S. labs: they haven’t spent much time on optimization because Nvidia has been aggressively shipping ever more capable systems that accommodate their needs. The route of least resistance has simply been to pay Nvidia. DeepSeek, however, just demonstrated that another route is available: heavy optimization can produce remarkable results on weaker hardware and with lower memory bandwidth; simply paying Nvidia more isn’t the only way to make better models.

https://stratechery.com/2025/deepseek-faq/

ngl I’m still confused

what the tech they made can actually do

It’s AI, it does AI things, is it because China can now do the things we do (coding/development/search queries etc) that are just as good as America that it’s a problem?

I’m slow, what’s the point? how does people joking about the fact China is censoring output explain

why the US tech sector is absolutely fucked going into the next generation

All that text to say not much? He doesn’t like that we want companies to pay tax

Fair enough, there’s no need to postulate what he might do when he’s just a moments notice away from threatening … literally anything, who knows what the crazy monkey will do or say at any time

Hard to tell at this stage but models may get a hell of a lot bigger if the hardware required to train them is much smaller or it may plateau for a while while everyone else works on training their stuff using significantly less power and hardware

Labor did nothing wrong, the voice was never ever ever going to pass unless we all got brain damage and became greens voters

Using mistral to calculate this but looks close enough

1921 + 1673 + 242 + 85 + 84 + 84 + 83 + 82 + 81 + 80 + 79 + 77 + 77 + 75 + 74 + 73 + 72 + 71 + 70 + 69 + 68 + 68 + 67 + 66 + 65 + 64 + 63 + 62 + 61 + 60 + 59 + 58 + 57 + 56 + 55 + 54 + 53 + 52 + 52 + 51 + 50 + 48 + 48 + 47 + 46 + 45 + 44 + 43 + 41 + 21 + 18 = 7264ms

So, the total sum of all the time values is 7264 milliseconds, or 7.264 seconds

Removing the snapd services: 7264ms - 3836ms = 3428ms

Criminal or not the golden rule has been around for a long time

So have wars over land

Their actions were and are immoral, dare I say evil

Welcome to the world, there’s a reason we signed up to a 300 billion dollar deal for some submarines and it’s not because we enjoy the look of them

Not really noticed it tbh:

1.951s snapd.seeded.service

1.673s snapd.service

Seems to be a lot longer than the mounts themselves but even then pretty minimal impact:

$ systemd-analyze blame | grep snap | grep mount
   86ms snap-bare-5.mount
   85ms snap-blanket-49.mount
   84ms snap-btop-813.mount
   83ms snap-btop-814.mount
   82ms snap-chromium-3010.mount
   82ms snap-chromium-3025.mount
   80ms snap-core18-2829.mount
   79ms snap-core18-2846.mount
   78ms snap-core20-2379.mount
   77ms snap-core20-2434.mount
   76ms snap-core22-1663.mount
   76ms snap-core22-1722.mount
   75ms snap-core24-609.mount
   74ms snap-core24-716.mount
   73ms snap-cups-1067.mount
   72ms snap-firefox-5600.mount
   71ms snap-firefox-5647.mount
   70ms snap-firmware\x2dupdater-127.mount
   69ms snap-firmware\x2dupdater-147.mount
   68ms snap-gnome\x2d3\x2d28\x2d1804-198.mount
   67ms snap-gnome\x2d3\x2d38\x2d2004-140.mount
   66ms snap-gnome\x2d3\x2d38\x2d2004-143.mount
   65ms snap-gnome\x2d42\x2d2204-172.mount
   64ms snap-gnome\x2d42\x2d2204-176.mount
   63ms snap-gnome\x2d46\x2d2404-66.mount
   62ms snap-gnome\x2d46\x2d2404-77.mount
   61ms snap-gtk\x2dcommon\x2dthemes-1534.mount
   60ms snap-gtk\x2dcommon\x2dthemes-1535.mount
   59ms snap-libreoffice-330.mount
   58ms snap-libreoffice-334.mount
   57ms snap-mesa\x2d2404-143.mount
   56ms snap-mesa\x2d2404-44.mount
   55ms snap-nvtop-171.mount
   54ms snap-pinta-33.mount
   53ms snap-pinta-37.mount
   52ms snap-snap\x2dstore-1244.mount
   51ms snap-snap\x2dstore-1248.mount
   50ms snap-snapd-23258.mount
   49ms snap-snapd-23545.mount
   48ms snap-snapd\x2ddesktop\x2dintegration-247.mount
   47ms snap-snapd\x2ddesktop\x2dintegration-253.mount
   46ms snap-surfshark-51.mount
   44ms snap-telegram\x2ddesktop-6489.mount
   43ms snap-transmission-100.mount
   42ms snap-youtube\x2ddl-4630.mount
   41ms snap-youtube\x2ddl-4806.mount
   40ms var-snap-firefox-common-host\x2dhunspell.mount
   24ms snap-telegram\x2ddesktop-6495.mount

Why not? That was 10 years ago and they’re trying to be profitable in a space that is extremely difficult to make profit.

each snap installed slows down your boot time

How so? How much impact are we talking? First I’ve heard of this

10+ year Ubuntu user here, how have I been screwed again?