Ask HN: Do regular users care about AI-enabled features? | Hacker News Reader

Ask HN: Do regular users care about AI-enabled features?

4 pointsQuadrupleA4 comments

It's clear that "AI" is being shoved into everything even vaguely technical these days. Every web app gets a new little magic wand icon to convert a pithy high-signal sentence into 7 bland high-noise sentences, or a Mona Lisa icon to more easily insert DALL-E images into your chats, spreadsheets, and fast food orders.

For me personally, as someone who uses LLMs daily for research and learning - and who has hooked LLM APIs to REPLs, my filesystem, made dozens of little specialized tools - I still barely ever use anything beyond the basic ChatGPT style interface. Smarter models help, and while I dabble with a menagerie of local LLMs I find myself still gravitating to OpenAIs flagship models, especially with the price drops of 4-turbo and 4o.

Anyway all that to ask - pardon my French - is there a world of consumers I'm not seeing that gives a flying shit about all these magic wand and Mona Lisa icons? Not talking about TechCrunch or the business press, or shareholders speculating on the impact of Google Gemini 1.6 Gamma Multimodal LabelMaker 6-Sigma Assistant, but actual everyday users?

Is this all a crazy FOMO gold rush set off by the phenomenal uptake of ChatGPT a few years ago?

What recent AI products beside ChatGPT-style QA and diffusion image generation (and github copilot for developers) have made an impact on the average person's heart and mind, and seen a lot of adoption and enthusiasm?

I wonder the same thing every day.

As a tech-literate user of the basic ChatGPT interface for the occasional coding assistance in an language or framework that I'm unfamiliar with, I pretty much *never* use it elsewhere. Nor do I miss it or wish it was integrated.

I see very few people asking for it to added, and plenty more people asking how to remove it where it was added.

I used a little image generation, but the process of trying to get an LLM to edit a generated image via a text interface is more work than it's worth.

I am in the States and have noticed the stars beneath each post on LinkedIn. Recently, there was a post about someone receiving their degree, and the AI suggested, "Would you like to learn more about XYZ University's degree?" As a mechanical engineer, I would not be interested in a master's in cybersecurity—it's completely irrelevant. I don't mind AI, but it should at least be relevant to the user. That's not too much to ask. Additionally, I've observed significant performance lags. What used to load in a second now takes several seconds. I'm not sure if it's just me, but it feels like a step backward. This isn't new to tech though. Remember how we completely moved away from plaintext emails to now fancier ones that contain a lot of visual garbage and links but not actual content?

It's going to turn out exactly like virtual assistants like Cortana and Siri where there's a massive rush to integrate it into everything and scoop up that VC funding (and hey, get that bag!) Because dorks think it's cool and maybe 10% of people actually use it and then 5 years later they have to rip it out of everything and it's on to the next grift.

They want to integrate it into desktop PCs? Lel. I did some training courses grad school classes and a lot of kids do literally everything on phones and tablets and maybe a chrome book. They use chatgpt to cheat and write papers because it's easy but nobody is using it for the kind of things starry eyed dorks are imagining using it for.

Nobody actually wants this except nerds and VCs trying to find the next gold rush. It's fancy spellcheck/suggestion/guessing and nobody gives a duck. I work with normies and half my team doesn't even have a modern enough smart phone to install recent apps. One guy just has a flip phone because he doesn't want to be bothered.

Remember how SV was pushing the Metaverse a couple years ago and now we all wear headsets and work in our Metaverse office using our NFT avatars with no legs? Exactly.

Silicon Valley isn't even worried about solving problems anymore except "the problem is I don't have enough money and I'd like to fix that". Nobody outside San Francisco and Austin is yearning for a fancy chatbot.

It might become more widely used when it’s integrated into commonly used software. For example, if Siri has an option for it, or it becomes the default option for Google. I don’t think regular users seek it out, though.

Reply on news.ycombinator.com