Who needs a ChatGPT Sports Car When An Alpaca can get you there for cheap?

I think my title worked better than my generative art attempt. But with all the slick new features in ChatGPT-4 being offered for paying customers only, and the suggestion that the costs of running Large Language Models is vast, would you like to hear of a bargain basement AI named Alpaca?


DALL•E generates me a disappointingly bad image based on the prompt “A llama riding in a cheap economy car trying to catch up to a rabbit in a red convertible, digital art” image placed into the public domain because no one really knows how to source these things.

Check this out:

Six months ago, only researchers and boffins were following the development of large language models. But ChatGPT’s launch late last year sent a rocket up humanity’s backside: machines are now able to communicate in a way pretty much indistinguishable from humans. They’re able to write text and even programming code across a dizzying array of subject areas in seconds, often of a very high standard. They’re improving at a meteoric rate, as the launch of GPT-4 illustrates, and they stand to fundamentally transform human society like few other technologies could, by potentially automating a range of job tasks – particularly among white-collar workers – people might previously have thought of as impossible.

But what about a language model you can build yourself for 600 bucks? A team of Stanford researchers has done just that, and its impressive performance highlights just how quickly this entire sector, and its awesome capabilities, might rapidly spin out of control.

Read the full story of how this low end Alpaca is trying to race against the LLaMA’s in their corvettes.

3 Likes

A little off topic, I think, how would you cite something generated by AI?

“Citing” can mean many things, and you can see a scramble as institutions try to develop guidelines for generated text and in there you might find MLA/APA recommendations.

I have spent a good amount of energy trying to look at how one might attribute AI generated imagery (do not expect anything authoritative) aiming for the Creative Commons elements of TASL- title, author, source, license when actually you rarely get 4 of these-- especially as the providers keep changing their terms and features.

Look at the approaches Creative Commons has done in attributing their use of AI generated imagery (I have been trying this) where it includes the platform, the prompt used, and some assertion of preferred sharing.

But it’s all fast moving, and frankly, for publishing now outside of my venues of blog posts and online writing, I might steer away from these images.

There is more than enough clearly open licensed media in services like Openverse– I can think the only case where AI images might be needed as to talk about AI as a topic.

That’s just my three cents. It’s moving too fast and wildly to find fixed rules.