Hacker News Reader

Comment by tantalor | Hacker News Reader

It says a lot about their product vision and intended market that the example query is typescript migration question.

Do they not care about typical search users? Only developers?

Compared to OpenAI, who seem keen to maintain the mindshare of everyone, IMO Anthropic are far more considered about their audience. They released a report recently on who who was using AI professionally and it was something like 40% developers, and single digit percentage for basically every other profession. I think they’re focusing on the professional use cases.

Pretty much. Claude from their announcements seems to me at least to be about SWE's and coding at the moment. Personally while I understand their decision I find it a bit limiting, and just a little targeted against the SWE profession. If all AI does is disrupt SWE's but not really add new products and/or new possibilities; then it feels IMO like a bit of a waste and is quite uneven in its society disruption.

At least in my circle SWE's are either excited or completely fearful of the new technology; and every other profession feels like it is just hype and hasn't really changed anything. They've tried it sure; but it didn't really have the data to help with even simpler domain's than SWE. Anecdotally I've had the comment from people around me - my easy {insert job here} will last longer than your tech job from many people I know from both white and blue collar workers. Its definitely reduced the respect for SWE's in general at least where I'm located.

I would like to see improvements in people's quality of life and new possibilities/frontiers from the technology, not just "more efficiencies" and disruption. It feels like there's a lack of imagination with the tech.

I know people in other industries use AI a lot and likes it. Accounting, legal, writing (a lot here). I agree that companies that focus on all verticals like openai is definitely the way to go. Claude code capabilities are not very significant compared to openai though. There is no big moat and a lot of it is perception, marketing.

Do users pay for LLMs? I haven't seen much concrete data indicating that they do. I don't think the casual utility gains of LLMs have gotten average people so much value that they're paying $20/mo+ for it. Certainly not for search in the age of [doom] scrolling.

I would guess that Anthropic wants developers talking about how good Claude is in their company Slack channels. That's the smart thing to do.

Have you actually done any kind of study on the utility the 'average user' has received, or is this just guessing?

I have only anecdotal data from non-technical friends and family.

I’m referring to average people who may not be average users because they’re barely using LLMs in the first place, if at all.

They have maybe tried ChatGPT a few times to generate some silly stories, and maybe come back to it once or twice a month for a question or two, but that’s it.

We’re all colored by our bubbles, and that’s not a study, but it’s something.

For most people AI is stuck at GPT 4 and other on par performance wise models. Anecdotally as well, many people that I know that have tried it found it mildly useful, but experience what coders and other tech workers experienced two years or so ago. Lots of hallucinations, lack of context, knowledge, etc. If you went back to those models you would at best feel like it is just a occasional code helper as well; at best an autocomplete or rather.

A lot of the reasoning model improvements of late are in domains where RL, RLHF and other techniques can be both used and verified with data and training; in particular coding and math as "easy targets" either due to their determinism or domain knowledge of the implementers. Hence it has been quite disruptive to those industries (e.g. AI people know and do a lot of software). I've heard a lot of comments in my circles from other people saying they don't want AI to have the data/context/etc in order to protect their company/job/etc (i.e. their economic moat/value). They look at coding and don't want that to be them - if coding is that hard and it can get automated like that imagine my job.

I would say no. While I pay for chatgpt Claude and perplexity monthly (I don't know why anymore) my wife does not use any at all. She has around 5-10 things she uses on the smartphone, and if she needs something new there is still google.

I on the other side reduced my googling by 95%

I'd guess they showed that query because LLMs are a lot better at answering translation/migration type stuff without hallucinating too much.

That's because the attention mechanism was designed for Seq2Seq models (i.e. translation in its most general form).

Any other use of it is a case of "I have a hammer, so that's a nail".

Reply on news.ycombinator.com