169
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 27 Aug 2026
169 points (94.2% liked)
Technology
87597 readers
5230 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 3 years ago
MODERATORS
"AI for the first time can replace or exceed human cognition." 🤣
No.
Maybe if you chose one tiny sliver of human cognition like memory and isolate it, but cognition overall? No chance. LLM tech is a database inference machine. Some uses? Yes. 1/3 of the worlds stock market wealth invested in it? Dumb.
Not many years ago, Bill Gates said that climate change was the most important issue facing the world.
I guess that's solved now, because he now says how we handle AI use and access has taken the mantle as the most important issue the world's facing.
This is just more AI cooker stuff from a billionaire octagonarian.
Who is so disconnected from reality he doesn't know the cost of groceries.
Oh he knows them. To him it is basically zero.
His net worth is 100.000.000.000$. The median persons net worth peaks at about 400.000$.
So in comparison what buying Groceries for the average person is when spending 100$; for him is equivalent to the average persons buying groceries spending 0,04¢.
He would have to go shopping 25 times to spend the equivalent of a cent.
But if we're playing that game then computers exceeded human cognition 80 years ago when ENIAC replaced human computers at doing arithmetic. If you count mechanical computers then it's 1000 years ago when mechanical tools were used to perform astronomy calculations.
Generative AI hasn't produced anything that exceeds human capability in any field. It's probably impossible for it to exceed human capability, since it's a statistical model of human output.
Chess is an early and canonical example of ai doing exactly that. The speed at which it can complete human-like tasks far exceeds human capability. It's hard to talk about the enormous problems with ai if we accept it as an undifferentiated point in the march of technological progress.
I think (comically) Gates uses the word "cognition" to mean, the business output of an individual white collar worker. In that realm, today, ais are faster, more durable, more scalable, more easily put into parallel than humans, they work around the clock and they're cheaper.
I don't like ai. Still, I think Gates is absolutely right that ai is going to swing a wrecking ball into white collar work and there's no government or industry plan to deal with the human impacts of that.
I hear you, but i disagree - at least based on what we know presently. If you look at US unemployment stats (which are a bit fucked from Covid), they jump in 2023 a little, but have since been decreasing each year since (employment is increasing). You can argue it's because the Trump admin is cooking the books in hwo they report unemployment, but an almost identical trend is seen in the UK stats.. Canada.. Germany.. Italy (barely even an AI blip in their stats)..
Are they all cooking the books?
If you believe the AI hype crew, AI is "improving in leaps and bounds weekly", but that simply doesn't align with reality.
Yeah, LLM AIs (the predominant AI tech in the AI© industry) have increased in quality as the sum total of all earth's investor-class capital is poured into them... But that isn't being translated in significant real world applications and job losses.
Companies tried (and are trying) AI©, and have mostly found it just isn't any better or more reliable than a human in the vast majority of real world applications.
If you are a white collar worker and you only make a few minor mistakes now and then - boss doesn't care, you're human. On the other hand if you're a white collar worker and you make only a couple mistakes now and then but wow they are batshit crazy ones that defy your coworkers attempts to understand them - you're fired. That's AI.
I think there's a reasonable, rational position somewhere between the ai hype cringelords on the one hand, and the egotistical humanists on the other who refuse to believe that ai could ever match human output. I think it's hard to look at what has happened over the past 5 years and have any confidence that ai will be constrained to some trivial, manageable ceiling.
I understand the skepticism people have around claims of the transformative power of ai. Some of those claims are just companies trumpeting their own products, and these breathless tech articles about how this or that tool is going to change everything, it's tiring. But I think that can numb us to the signs that something transformative and irrevocable is actually happening.
Even if we assume that current ai is at its limit of progress (it isn't), if an agent costs $10k/year and works 24 hours per day, there is a far greater tolerance for mistakes than a worker who costs $50k and wants vacations and benefits, and sleep. Gates makes this point, and I think it's strong.
Layoffs due to ai replacing human work are happening in the tech sector already, which I think is arguably a canary in the coal mine situation.
You're begging the question.
How so?
https://en.wikipedia.org/wiki/Begging_the_question