Such a cool guide to such a bizarre problem we live in…
Companies forcing features that nobody wants, but that are also expensive to operate and cost real money to run? Market can definitely stay irrational for long.
Companies forcing features that nobody wants, but that are also expensive to operate and cost real money to run? Market can definitely stay irrational for long.
By then, a new fad will come in and you can just pivot to it.
We went from "engagement" (with no concrete plans to monetize said engagement), through a brief period of "blockchain" and now finally AI.
Google and Meta have made billions monetizing said engagement
Imagine you make a widely hated product targeted at huge corporations. Workday, Jira, Oracle Financials, anything like that.
Then you find that all your users are using ChatGPT to automate their interactions with your product. Instead of your product being crucial to your customers' business and almost impossible to replace, it's turning into a dumb pipe connecting LLMs - and looking very much threatened.
After all, data lock-in is a core part of your business strategy.
Getting people to use your AI tool, though? That makes your product stickier, and lets you charge more too.
> That makes your product stickier, and lets you charge more too.
Literally every day that I browse HN I hear people complain about others' use of LLMs. How can you say "nobody wants" when it's used so much to elicit endless complaints?
That’s what I’m talking about, not AI use in general.
I suspect that many people dislike AI but are begrudgingly using it. I would put money very few people are completely avoiding it. It's very hard to avoid now anyways.
Maybe the AI company strategy is really just "make it so invasive in life it's so difficult to avoid and expensive enough to revert that companies and governments will never go back anyways"
(Sadly,) many normal folks use ChatGPT in lieu of Google search. I'm continually surprised to, over and over, hear from personal friends that they planned their vacation using ChatGPT.
They were not begrudgingly using it.
If it's popping up in some random product I don't pay for I can only assume it's such a shitty model that it's probably not worth talking to.
(Exception: my bank has replaced its shitty phone tree with an AI chatbot. This is faster and the model has a "transfer to a human" tool. It's perfectly capable of following its playbook, determining that I need a human, and putting me in touch with one. So actually I started with "now I've gotta debase myself by talking to a fucking 32B Chinese LLM" but ended with "oh that was great actually").
It's just that another group of people absolutely hates the idea of AI and wants it to go away.
The simple distinction of pro vs anti ai isn't the root cause, its about individual control and choice which can span both of those stances.
> I think people like AI tools when it's their choice to engage with them and they set the terms.
Every AI phone assistant I've talked to has been pretty terrible, even compared to modern LLMs like ChatGPT voice mode.
I mean, many of the features had previous pre-AI attempts. They all sucked much more. This is the first time they have a remote chance of working (but in many cases, timidity or cheapening out is still making them barely usable - which is why 99% of the cases it's still much, much better to throw one of the three SOTA LLMs at the app to operate it (or on its data) from the outside).
The LLMs will always do the worse option to customers, to extract more money. Burn'em all to the ground.
Shitty chatbots became a thing a few years before LLMs. The LLM variant is an improvement for users - it's more typo-tolerant, and more likely to understand what the user actually wants, before it flat out refuses because of organizational policies.
> The LLMs will always do the worse option to customers, to extract more money. Burn'em all to the ground.
That's a ridiculous take. The models are very capable. What they actually do depends entirely on constraints set up by the companies deploying it. If the AI features on some site "always do the worse option to customers", it's not a problem with LLMs - there is a human being in the company, that choose to fuck over their customers.
Same is true about pre-LLM support. It's not people from India who are at fault for your shitty support - it's the executives in the Western company that explicitly decided to outsource their customer support to a call center on a different continent, and they're also the ones who decided that "terrible scripting and half-broken english" is the level of support they will offer to their customers.
Same with manufacturing - "Made in China" is a synonym of bad quality, but it's not because of China. The brands are free to seek out any combination of quality and price, and it's them who choose to dial quality down to minimum. Factories don't care what's in the spec, they do work to spec.
Same with LLM chatbots. Prompt them to be unhelpful and prevent them from accessing important information or paging actual people, and they will be useless. Not the fault of LLM.
Yes, it's my pet peeve, and it's been such long before LLMs. There's a pattern of blaming technology and technologists for shitty or harmful work, while actively ignoring the business side that devised and commissioned it on purpose. Stop blaming LLMs. Start blaming entrepreneurs who chose how to deploy them.
The pro-ai posts here have got to be 50% bots, i'm convinced. I have yet to see a single artefact of AI that was remotely impressive. Fancy markov chains do not intelligence make.
The bigger problem with these apps is that they are proprietary and they have been spying on us.
The article refers to different AI features within other apps (e.g. search overviews in google, summarize email, summarize message) that bring no benefit to the user but only clutter.
I think everyone is fine with using their AI of choice to do their tasks of choice. It’s when I see a button that obstructs my workflow prompting me to do something I don’t need to with a shitty model that’s not good at it.
If I personally wanted to use generative AI, I’d query an agent app on my computer or load ChatGPT in a browser. I wouldn’t want my email client proactively spewing random words into the text area on its own. And I certainly don’t want every bit of information on my screen continuously streamed to a corporate cloud server so that I can ultimately be profiled and surveilled.
Look at OpenAI revenue just over a year ago. Most of that was not for LLM coding.
Loads of non-technical people I know are paid subscribers $20/mo to ChatGPT (not API, but their actual chat product).