[pavlov]

I wonder if this gold rush is leaving openings for building simple solutions that don't need AI and emphasize traditional UX instead of conversational UI. If you have some domain knowledge, maybe you should focus on ensuring your product does a narrowly defined 100% job at what's needed, and it may shine when you're pitching against AI-based competitors whose products look like magic 95% of the time but are wrong or unusable the remaining 5%.

"We are not a black box that sends everything to someone else's giant API" can be a selling point as businesses become more aware of their cloud dependencies.

[kolinko]

The basic simple (web) solutions were built over the last 10-20 years - of course not everything was invented yet, but if there is a need for such a product, there is a high chance the product was already built, and the niche filled.

Perhaps a better question is what new niches now open up, on top of the AI solutions.

Hugging Face is a nice example - it's mostly a python library, a github clone, and a discussion forum strapped together. "Old" tech used to serve a new and growing niche.

[geraltofrivia]

HuggingFace is also hosting models (CLI/code/API), have an extensive community driven dataset store, and has some great positive feedback loops.

Great Library -> Large Community -> Large Coverage of Models & Datasets -> Larger Community -> Larger Coverage of Models & Datasets -> Revenue -> More Engineers -> Greater Library ...

Not all the steps (->) here are trivial: Contrast HF to Explosion (great folks behind spacy).

[banditelol]

> Not all the steps (->) here are trivial: Contrast HF to Explosion (great folks behind spacy).

Can you elaborate more on this? about contrasting HF to Explosion?

[geraltofrivia]

Explosion's core contribution (not moneymakers) is the great SpaCy library they first released in 2015. It was an excellent work, far better designed (IMO) than NLTK and other offerings at the time. Of course the library isn't monetized. SpaCy too has the ability to train custom models and use them.

This never transformed into a model hub. Despite a lot of people using SpaCy and probably building custom models.

Again in contrast, Explosion's other revenue stream (prodigy) is not a SaaS as well. Its a great software, and I presume it brings in a steady income. But in 2023, I would imagine that HF's LLM hosting, cloud training environment, brings in more money than Explosion's data annotation software.

[z3c0]

I'll also add that Explosion has a heavy "production-ready, quickly" bent, and even supports wrapping HF models with spaCy. Explosion is probably my favorite company in the AI space and has provided the most tangible value of all the NLP tools I've used.