[DoryMinh]

Fantastic, now we have duopoly

[ianhawes]

IMO the rankings of publicly available LLMs are:

1. GPT-4

2. Claude 2

3. Bard

4. Llama/Alpaca

5-98. [Unclaimed]

99. SmarterChild AIM bot

100. Cohere

All joking aside, I do agree with the sentiment that no one generally has any type of defensible moat at the moment. OpenAI has found a great balancing act between first mover advantage, marketing, customer adoption, and enterprise sales. They are executing at a high level. Anthropic (Claude) has a wonderful product but is lacking in consumer adoption and sales, though I think they're working on fixing that.

[ilrwbwrkhv]

All the AI companies are sort of doing a VC rush, but instead of IPO it's AGI. Would be fun to see what we get in the future. Since a serious training run costs upwards of $50 million currently.

[AviationAtom]

I think the reason Claude isn't further than it is now is due to a singular goal:

Do no evil

I feel they are what Google was to search engines when they burst onto the scene. They'd rather take time to get it right than to push out a product they don't feel meets it's mark.

Now that they have entered open beta I expect them to rapidly iterate. If their product is as good as I feel it already is, in comparison to competitors, then I can only imagine what it will be doing in a year's time.

(Disclaimer: I do have exposure to Anthropic through my investments, but nothing I said is based on that alone)

[usaar333]

Seems inferior to GPT-4 on every test I've given it - but as a competitor to GPT 3.5 is strong.

[abdullin]

On our benchmarks, Claude v1 beats GPT-3.5 (v0613) while v2 looses to it.

[alpark3]

> monopoly

[gberger]

There is no moat.