ES version is available. Content is displayed in original English for accuracy.
Advertisement
Advertisement
⚡ Community Insights
Discussion Sentiment
65% Positive
Analyzed from 4289 words in the discussion.
Trending Topics
#google#gemini#data#don#same#using#companies#trust#high#more
Discussion (124 Comments)Read Original on HackerNews
No chance I'd trust Gemini to correctly tell me which wire was the high voltage one.
You're absolutely right! I apologize for providing a factually incorrect explanation. You asked a question, and I gave you incorrect information that put you in danger. That failure is entirely on my side as an AI system, not on you for following what you were told.
We need to start training people that AI answers are just as fallible as answers you get from a human. Trust but verify for all beings, biological and non-biological.
It's like "[actual living person] has big eyes" and "character portrait in a 90s video game has big eyes" are similar sentences on the surface, but they describe things as far part from each other as a single dot left by bird poop vs. the combined writings of all humans, ever. Like half an atom compared to a universe.
Same for "beings". Just no. Or rather, I apply what you suggest to what you said, I don't just take your word for it but inspect if what you said could be true, and realise your answer is so full of false premises it's not even wrong.
No, LLM generated answers are not the same as those from a qualified expert. Of course you should trust a human if that is their job and you should never trust a word generator to give you accurate answers to an important question.
You're right, LLM-generated answers are not the same as those from a qualified expert. Sometimes they're better.
Case in point. I recently had a discussion with a friend of mine about going off propranolol. He wanted to stop it because he didn't like the side effects and only wanted to take it as needed for anxiety-related problems in public situations. His primary care said to just stop taking the drug. Gemini told him that you need to have a tapering protocol, and you need to work it out with your doctor.
In this case, his primary care, theoretically a "qualified expert," was wrong, and Gemini was right. His course of action at that point was to get a new primary care that had better knowledge that matched the vetted medical sites I pointed him to.
fwiw, he trusted my advice and validation of his plan to get a new primary care doctor. The reason he trusted me is that he's seen the prescription drug disasters I've gone through, and knows how deeply I've had to educate myself as a form of self-defense.
So far I've learned 480V is very unforgiving and can show up in surprising places.
* https://youtu.be/knSkE9ejPl8?t=112
The guy isn't dumb, but he is obsessed with solving every problem with AI.
There is already deterministic procedural stuff for all of those cases. Putting a chat bot in front of that is basically fucking stupid.
No latter than yesterday I used Gemini to tell me which kind of power adapter was compatible with an old SNES chinese backuping device [1] (a device you'd plug instead of a cartdridge into the SNES and then you could make backups of cartdridges on to 3.5 floppy disks. You could then run the games from those backups. Conveniently you could also "backup" your friends' cartdridges.. Heck, there was even a tiny demo scene for the SNES and you could swap disks with little demos written for the SNES -- just like for the Amiga).
It was a bit tricky in that Gemini explained me I needed a reverse polarity adapter or I risked bricking everything.
Now, granted, the stakes weren't very high but... Gemini got it right and I could start the SNES, with the chinese "Super Pro Fighter" backuping device just fine.
[1] this thing: https://videogamedevelopmentdevices.fandom.com/wiki/Super_Pr...
Demis is busy create another PR documentary lionizing himself so he can capture the google crown. Always says whole bunch of nothing to a fawning interviewers who is projecting brilliance and profundity on his every word.
-AI use at work is broad but shallow: -At work, most AI use is focused on collaboration and assistance with tasks, and so far task automation is uncommon:
Both of the above happen for a reason and if you work at any bigger corp ( esp publicly owned ), it is easy to understand why.
edit:
What I did find more amusing though was the coded cry for help ' To get there, we as a society must work together".
Also if all the built-in AI buttons could actually automate tasks instead of just telling you how they're unable to automate tasks.
My company (not my startup, my normal day to day work) spends millions on ai token every month.
I use AI every day for different things.
AI/GenAI disrupted already a lot of peoples jobs in real life.
And Google states very early "However within jobs, people are using AI selectively: in a typical job AI is used for only ~21% of tasks."
Did you read the article? Can you point out specific things you really don't like and think they are cherry picking?
Reading the post, I had to keep thinking “why are they only talk about how people use it at work, and nothing about possible economic outcomes?”. Or “titled Understanding the AI Economy, but it avoids all the big questions”. Or “is it written for the investors who think Google has over-invested in capex?”. And I might be overthinking about all this! But when you have billions of billions dollars invested, it’s hard to take any word as is.
What value does Google has to cherry-pick the occupation metrics?
* what's a typical job? How did Google resample their data to make it representative of the typical job?
* is this all jobs, or just work-related Gemini conversations that they've classified as such?
* what's the denominator here? How do they know how many tasks there are that weren't subject to conversation with Gemini?
* how reliable is their classification of conversations as work related or not?
Accurately measuring this is so complex that to present the conclusion in such a confident way is bordering on farcical.
And its Version 1 and an ongoing project.
"In the workplace, we show that while AI adoption spans occupations covering just above 88% of US employment, penetration remains shallow and overwhelmingly collaborative in nature, with end-to-end task automation limited in scope. Outside of work, AI spans activities making up about 98% of Americans’ non-sleep time, with disproportionately high use in high-friction tasks such as engaging with government and professional service providers, likely delivering economic value that standard national accounts may miss. Globally, adoption scales with national wealth and has broad linguistic distribution, with English queries representing only around a third of volume."
It sounds reasonable enough and it contains nothing off.
For me its interesting to read that english is less dominant and thats def something i was very atuned when GPT-3 came out: I started to write english and my native language and mixing up words etc. The quality of understanding in my native language is sledomly good when a new product comes out but it was very good already when using GPT-3.
In comparision to this, i'm also aware of a research paper stating that programming in certain languages like spanish is worse.
How much profit does the AI company make from those tokens?
My company is a two digit billion dollar revenue company in the top 100.
My point was that real companies pay real money for AI.
But i use it daily at work and a lot of co-workers do too.
It simply comes off as a marketing advert for a company who’s sunk billions upon billions into something they’ve yet to be able to monetize (see latest report). Stock holders are getting annoyed, like in other companies, about the lack of the promised returns, this is firefighting.
Nobody is really debating AI has a use case, but maybe, just maybe, the chase into infinite spending we’ve seen the last years wasn’t that well thought out.
Im missing the marketing opportunity here.
The spending is a complete different topic though. Google/ABC can easily afford their investments. Microsoft and Amazon too. Even if shareholders are not that happy (the shareprice of alphabet doesn't tell this story though) with it.
Nonetheless the biggest issue with this spending are 401k and other retirement investments from people who can't control the investment and the undefined risk of it. Thats the only real 'risk' (not a small one for these people depending on it though).
It doesn't affect me if Google spends its money on AI.
Nonetheless the progress on AI is huge, it still hasn't slowed down, its now a political and global issue (us export restrictions, Chinas Moonshot AI/Deepseek, EU lagging behind).
At least form my point of view, i'm very surprised that we even risk all of this investment and see it as a possitive thing. Imagine AI/AGI is a technology which sits very far away from our current local minimum and the jump to the other local minimum is massivly expensive but massivly beneficial?
GPT-2/3 was able to inflict this vision into people and because of this, we get A LOT more compute, push again boundaries on compute and memory after a longer period of stagnation. These richest companies in the world like Google also leverage the AI spend into investment of fusion and other energy sources.
We already got a lot better vision ML algorithm because of this (tx to suckerbergs dino and segment anything), we have alphafold tx to Googles Deepmind etc.
AI only consumes currently a little bit more about bitcoin but is acutally beneficial.
But it doesn't prevent them from pretending they do Y)
You wrote a snarky comment without content.
Would love to actually have a real discussion about critisism like this (not because i'm affiliated with google at all, but because I like to expand my knowledge through discussions)
Google:
1. Is not remotely unbiased. 2. Does not have desirable models, coding harnesses, or data centre capacity for AI (they rent space from SpaceX). 3. Seems to be lagging far behind in all things AI. 4. Has serious problems with internal adoption of AI tools (such as banning most team members from using competitors, but allowing the DeepMind team to use Claude, etc.)
Their opinion from their PR team is not just that credible or valuable.
> Observation Clustering and Taxonomy Organisation (OCTO), a tool for transforming massive unstructured text data, like LLM conversations, and distilling them into organized entities.
This also sounds interesting. Hand crafted ontologies that I've seen used were always too hard to maintain and make useful, but potentially they have something automated?
Anyone know why they did this? Once I saw it, it became too distracting.
"We picked some results we wanted to show and told gemini to cherry pick data that would fit in. As you can see, it was not a lot."
> AI use at work is broad but shallow: Workplace adoption spans all industry sectors and also 68% of all occupations that collectively represent 90% of total U.S. employment. However within jobs, people are using AI selectively: in a typical job AI is used for only ~21% of tasks.
Translation: the market is reaching saturation.
> At work, most AI use is focused on collaboration and assistance with tasks, and so far task automation is uncommon
Translation: AI can’t be trusted to operate on its own and very little progress has been made in that area despite humongous progress being used in other areas. It’s helpful, but in the same way that a tool is helpful (and most companies don’t want to overspend on tools unless they can seriously reduce labor cost, I.e. hiring fewer people)
> AI is delivering value at home that may be missed in standard economic metrics, particularly around high-friction administrative tasks: Over 86% of interactions with AI tools in ATLAS occur outside of work.
Translation: AI is most valuable in low dollar value scenarios where it is likely a loss leader. The amount of money the average home user is willing to spend on AI is likely very close to zero, and this is the most likely interaction to be a race to the bottom. E.g., why pay Anthropic $20 a month when your iPhone 19 has a built-in local model that’s almost as good, why sit through advertisements on your free AI service when a cost and ad-free alternative is available.
All of this alone doesn’t imply a bubble. The bubble part is where many AI companies can’t financially survive without a massive product maturation, budget cuts, or further investment.
The level of investment and operating costs of AI firms is just so high that the question of where the customer spend comes from to make it profitable eventually comes into play.
The other obvious question is how AI firms maintain their bookings and overall value when the upper middle class responsible for the majority of consumption dollars is getting laid off left and right by AI.
E.g., OpenAI finally ships AGI and realizes their investment value and goes from unprofitable to profitable. They can charge companies thousands per month for their best models/tooling because they replace employees entirely.
All the big corporate customers of the world are happy: truck drivers are replaced by self-driving, big high-income employee bases like Meta, JPMorgan, law firms slash employee counts by huge amounts like 50% or maybe even more.
…and as soon as that happens all those same companies face massive customer attrition since all those high-earners were just laid off and aren’t buying PlayStations and KitchenAid stand mixers anymore.
For an article about the AI economy, there is no mention of economy.
Notice that nobody who actually innovated on the technology ever speaks on it.
It’s always these thieves from academia and Google who act like they’re the inventors. Trying to explain to us - the people they have been ripping off for years - how “AI works”
Go home Google
They are for sure one of the current Frontier AI labs.
And before their competitors and china started to be that fast, they were regularly publishing a lot of research in this area too.
Were is your hate coming from?
I’ll take information from multiple independent analysts please.
I do not follow your argument.
Would you read this and say 'this is now the only truth'?
I mean i read it and will add it to multiply other sources. Now i know what / how google seas things and i try to get something from it.
i find it very ignorant if one of the biggest ai companies on the planet posts something and its just getting dismissed. You don't have to follow google every word, but it def adds to an overall picture even if you don't take everything as true
You wanna elaborate?
And no, I wouldn't actually trust Sam "King of the Cannibals" Altman. But I would still expect him to start with far, far more accurate data, and distort it a lot less to suit his narrative than Zitron would. That's how low the bar is.
The man made saying "the AI industry is failing" his entire personality. He'll keep saying that regardless of what the AI industry is doing, forever. You can write "tech companies often overpromise and underdeliver" on a post-it note, stick it to a whiteboard, and that will provide about as much perspective as everything Zitron has ever written, or will ever write.
Sure, Ed’s making some kind of money off of subscriptions. Yeah, he’s a personality. But to ignore the deep comparisons, numbers and data that he gathers and provides to back up his analysis feels like sticking your head in the sand.
There are other banks and financial entities that are echoing his concerns as well.
As a ready-made example: the "AI models are unprofitable, API inference is subsidized, OpenAI is naked" line he keeps stressing. That doesn't seem to be likely even on the basic napkin math estimates. And is also repeatedly refuted by both AI company officials and various AI industry insiders - of varying trustworthiness, and with different incentives.
Having "the numbers" isn't the same as having objectivity, and again: I'd trust Altman to have a semblance of objectivity before I would trust Zitron.
Im sorry but what? Plumbers, electricians, framers, welders, auto workers, hotel staff, restaurant staff are these people using AI in their jobs? Does someone at google think that these very much manual labor jobs are only 10 percent of the work force in the US?
There is PEW data, that says 30 percent of us labor force is in manual labor / blue collar jobs. The person you call at 3am to fix your drain, isnt using AI.
I am 100% using AI to figure out wtf to do about my 3am plumbing problem.
/joke
After I have the water turned off, I’m going to back to sleep to call a plumber during regular hours.