Right now GPT4 is useful/powerful and probably anything else sucks. That may change but the mountain has real value at the top only. At least for
chat used for either human assist or decision making or code generation. Categorisation, yeah you can use lesser LLMs.
Not necessarily true anymore. As a counterexample, consider CodeLlama 33b, which is quite good (and which has replaced GPT-4 for my coding assistant needs).
OpenAI's models are likely to remain the best, but I see open models catching up and becoming "good enough." Why pay for GPT-4 when I can run a model locally for free? (Barring the initial capital cost of a GPU; and not even this if you're, say, using a Macbook)
Like yes, as long as GPT-4 remains the most useful (and now they have multi-modality in Plus so they've just extended their lead) nobody will adopt anything else.
So grandparent's argument makes even less sense as search should also be fungible but it isn't when Google still provides the best search product and has since it started. If OpenAI can maintain that lead and continue to ship improvements that continue to push their peak higher, most people will not pay for anything less even when that price is free.
I don't work in tech and I use LLaMa. It is great for giving me ideas, writing overviews which I use to make sure I have most of my bases covered (and helping me see which of my ideas are more original), getting more examples, along with other creative tasks.
Could I get my employer to pay for it if all free options disappeared? Probably, but I don't have to while they exist.
