Live data from Hacker News

Not everyone is using AI for everything

gabrielweinberg.com

481–490 of 565 posts

Re: Not everyone is using AI for everything

#481
post #82
post #72

On the post-grad job hunt right now - I note that most employers will ask in a technical interview or whiteboard interview "how are you using LLMs?" It's tough to answer because you want to hedge for both an AI enthused employer and an AI hesitant employer with limited information about who they are and how they personally use these products. I've been responding with a sort of long winded answer about how 'there is…

> It's tough to answer because you want to hedge for both an AI enthused employer and an AI hesitant employer with limited information about who they are and how they personally use these products. Have you considered just answering truthfully? Would you even want to work somewhere where you need to play a role and where they flip out when you say the wrong word you should've correctly guessed through mind reading? T…

Interviews work both ways: employers select employees, employees select employers.

Unfortunately often employees forget that part of the goals of an interview. An employer never forgets its goal.

Of course the asymmetry stems principally from who pais whom and secondarily from who can wait more to find a better fit. Sometimes that's the employee.

If one is not desperate to get a salary immediately, it helps thinking that the company applied for an interview and agreed to let the interviewer come and inspect their premises, their personnel and their processes. The right questions derive from that realization. By the way, that's what auditors do.

Re: Not everyone is using AI for everything

#482

It's important to factor in just how many US adults are basically illiterate nowadays. As of 2023, 27% of American working-age adults were at a PIAAC Literacy Level of 1 or below, out of a total of 5 levels. This has gotten drastically worse in the past 10 years as, in 2013, Level 1 and below was only 17%. Full scores for 2023 are: % Level 1 or below: 27% Level 2: 29% Level 3: 31% Level 4/5: 13% For reference, Level…

"Response rates for this data collection were relatively low, both for the United States and for several other participating countries. There is evidence that procedures implemented to reduce bias associated with nonresponse have done so, and that the data are representative of the population. However, readers should be aware of the potential for bias and use caution when interpreting PIAAC results." These stats don'…

> About a third of people in the US have a bachelor's degree, but only 13% can pass level 4/5 literacy challenge?

Yes, that's how bad the problem is.

I've worked with people with master's degrees who can't comprehend text more complex than a Slack message. I've watched them try and fail.

I think it is difficult for a highly-literate person to wrap their head around this. It's very possible to get through a bachelor's program with low literacy. It's even possible to do quite well. And at a reasonably good school.

There are some pretty successful businesses out there whose target market are essentially professionals with low literacy. They use more flattering descriptions in their marketing.

Re: Not everyone is using AI for everything

#483
Almost everyone is using an app with a feed with an algorithm powered by a transformer. People spend hours a day feeding information into a model with the same architecture as an LLM which then spits out next-post recommendations. They call it "scrolling Reels/TikTok/Shorts".

That has to count for something.

Re: Not everyone is using AI for everything

#484

Earlier quoted context omitted.

I assume it's because he is seeking to pay rent, food bills, and other expenses through employment.

But it doesn't make sense to do a job you'd hate. Just be honest, land the job where you're happy.

> it doesn't make sense to do a job you'd hate

That assumes using/not using ai will make me hate the job. Which is not true at all for me. Would be perfectly happy either way. (Or rather to say my happyness would depend on other factors.) Obviously i wouldn’t want to be the only team member not using AI in a team using AI or vice versa.

Re: Not everyone is using AI for everything

#485
post #367

Good piece, but I think there's a missing angle to it. He cites a study showing how often people say they "use AI", and a little over 50% use it less than once per week. If we're just talking about AI chat interfaces, sure. But I think the way that AI usage is going to grow isn't mostly by getting more chat engagement. It's about baking AI features into software that people already use. For example, suppose you asked…

I'd say this argument is not relevant to the specific question the article tries to answer, as AI adoption through these means is forced and may in many cases go against user preferences. Edit: The deciding factor being whether you want to figure out if people are interested in AI / find it useful, or if the question you seek answered is more akin to "X% of people consume lead in their food"

People eating lead in their food don't know it's there.

People engaging with the AI built into Google results pages can see it, and assumedly Google's A/B testing showed that they engaged with it.

Re: Not everyone is using AI for everything

#486

Earlier quoted context omitted.

“I assume it's because he is seeking to pay rent, food bills, and other expenses through employment.” Fair enough, so if there were one “right” answer, that would be the one to give whether true or not. But here there is no obvious right answer. If the employer is looking for a particular answer, the poster doesn’t know what it is. In that case, the best thing to say is simply the truth, particularly when the truth t…

The best thing to say when you don't know the answer probabilistically is to give the most likely correct answer.

Common sense not common

Re: Not everyone is using AI for everything

#487
post #383

Earlier quoted context omitted.

The behavior of an LLM is not and cannot be “well-specified”.

trumps voice wrong https://developers.openai.com/api/docs/guides/structured-out... https://github.com/guidance-ai/guidance https://github.com/noamgat/lm-format-enforcer https://github.com/mlc-ai/xgrammar https://github.com/dottxt-ai/outlines

Libraries like Guidance guarantee that the output of an LLM will be syntactically correct (i.e. it will be valid JSON or whatever output format you are wanting). They do not, and fundamentally cannot guarantee that the data contained in them is actually correct, and cannot make the actual behavior of the LLM "well-specified". Or, as you put it, trumps voice wrong.

Re: Not everyone is using AI for everything

#489
post #239

I think the gap is because 1. For coding, Claude is amazing - mainly because of its curated skills and because massive amounts of working code has already been carefully labeled over the last decade or so via GitHub. And because with any Turing complete language, there is only so much one can do. But 2. For most other things, LLMs are fairly underwhelming. Research is usually mediocre. Try being rigorous and repeat y…

Basically this. For tax, and law in general, LLMs are prone to giving the wrong answer for anything specific (though they're great at the super general stuff because they're just restating one of the millions of summaries written by humans).

Gemini still isn't sure what details are in the version of OBBA that actually passed, because there was more discussion about various proposals (that didn't make it into the bill) than there was about the final version of the bill itself.

Unfortunately, it's an intractable problem based on the ways that LLMs work. In order to overcome those limitations, you have to provide so much detail to the prompt that you would find the answer faster searching manually.

Re: Not everyone is using AI for everything

#490
post #69

Some of the advantages are second order. For example; ChatGPT is replacing my Google searching. Not necessarily because it's better, or because it's summaries are better than Google (I find them subjectively better but it's not clear cut). But because the app has a nice history; can ask a relatively complicated question and go do something else and then come back to it, ask a follow up. Etc. None of that is specifica…

That's funny, Google Gemini and AI mode in search has replaced my ChatGPT prompting, because I know Gemini will correctly cite sources (as of course it's by Google) rather than hallucinating. Also, Gemini is free or at least has much higher usage limits than ChatGPT or Claude, and it's well integrated into Android and soon Apple with their new Siri, so things like circle to search just work well.

I've run into plenty of instances where Gemini cited sources and the claims they attributed to those citations were nowhere to be found in the actual sources they cited. It's happened to me over a dozen times this past week.

The AI search results were actually better last year; significantly fewer hallucinations.

Post reply on HN