Hacker News Reader

Comment by ulrikhansen54 | Hacker News Reader

There are several stories of companies going from $0 to >$5M ARR in just a few months with the boom in LLM-driven applications. The downside for these types of apps is that churn rates are through the roof, and will probs die when the LLM fad inevitably fizzles.

View on HN

LLMs aren't a fad though. Yasser has thousands of customers paying him real money because they find value in what his product does.

I'm running a somewhat similar website (https://Docalysis.com/) where users chat with files, and it's clear there's a lot of value being added, so people are willing to pay.

What's less clear is how this all plays out when there's more competition, but it's not like we'll all go out of business. It'll just be a bit harder, or you'll have to do things to differentiate. I'm planning on differentiating more, just using the current product as a starting point. Yasser probably is thinking along similar lines.

Beware that most 2020-2021 OpenAI startups haven't survived to today, especially since ChatGPT allowed free use of most B2C use cases (article writing, paraphrasing, etc). I posted a brief postmortem of my experience from 2021 to 2023 when I did at one point launch a similar product to yours which, if you got into this more recently, may be worth reading.

I do expect constraints to emerge, in the form of regulatory capture and copyright challenges from content owners.

The bigger play is maybe data lakes. Upwards of 80% of that data is unstructured. And, it likely can’t move off prem.

New architectures and models will inevitably replace LLMs, and it'll probably happen pretty quickly. The models we're using today will be outdated in 12 months time. The next big thing will a fusion of LLMs, LVMs (large vision models) and LAMs (large audio models) into modal-modal models.

Like any gold rush? Right now, gold nuggets are strewn around and you can grab them with your bare hands. Pretty soon you'll need excavators, sifters and all that infrastructure.

Good on him for making this much cash. Sure, right time and right place but you still need to be able to execute.

Of course, it will not last forever but like being "devops" when AWS started gaining traction or jumping on the crypto bandwagon - now you know how to play the game, there's literally no downside to jumping on this bandwagon if you can find a profitable niche even for a while as you gain expertise.

My company is considered 'late moving' ; we like stability, no stress, high profits, healthy growth (no growth is better than unhealthy growth) etc. We are at almost 2 decades of profit (and a lot of profit at that), no VC capital or other capital for that matter.

We have been adding LLMs to our products in a little test for clients who want to try it. It's not a fad as far as I got the feedback; it saves people are lot of work and allows juniors at our clients to do far more. I believe it retains clients and will pick up new ones. So when the fads with LLMs die out, the products that have marketshare and add them (b2b anyway), will benefit a lot.

Why would it fizzle? It provides an actual utility

There are a lot of companies "embracing AI" for CEO-fell-for-the-hype / CEO-hopes-investors-will-fall-for-the-hype reasons, that aren't related to actual business needs or benefits from using AI. Lying (to themselves, and/or others) about how use of AI is improving business efficiency. Forcing it on people down the chain whose reality-based "um, this mandate to use AI tools for this is actually making things take longer" complaints are being ignored.

Many of these companies are throwing money at these kinds of early-mover products, but are likely to stop doing that as people start to get wise to just how useful all this actually is, and for what.

[EDIT] This is not to claim that AI-based products won't find actually beneficial uses in various companies, but a lot of these trivial get-to-market-fast companies are raking in quite a bit of money from hype-chasers, and that money isn't going to stick around long-term.

I think it probably won't fizzle, but one reason it could is if competition drives down margins to the point where LLMs become a commodity while technical progress stagnates and their abilities saturate. Another possibility is that they could get banned or heavily regulated.

Reply on news.ycombinator.com