I made TypeSafe AI 's new ultra fast model, Jev, generate text, even though it shouldn't, that's fine because I can't read, and it can't write** ** up to 20 words for 0.5$, what a steal

@typesafeai How to turn a cheap model into the most expensive af model

@typesafeai That is one way to break it xD

@typesafeai LLM from first principles

@typesafeai 147k tokens to write a para that gives up halfway through is hilarious.

@typesafeai i saw someone make a diffusion model using jev too. it was also garbage

@typesafeai Cool stuff. The demo is amazing. I also built this recently

@typesafeai JewGPT please 🙏🙏🙏

@typesafeai lmfao

@typesafeai that made me laugh🤣

@typesafeai Lmfao

@typesafeai did you try the mobile repo where they included Jev

@typesafeai back in the day we just spammed the next suggested word on our phone keyboards

@typesafeai is this ... is this ... a GPT-simulator ? 😀 😉

@typesafeai Truly amazing work 😆

@typesafeai tried it too , not really good , generated a lot of goober

@typesafeai this is the way.

@typesafeai I get the humor, but charging for low-quality output feels a bit sketchy

@typesafeai It is definitely built for structured decision-making rather than chat, but the speed is incredible. We actually went deeper on this here:

@typesafeai that would be chatJev GPT is for generative pretrained transformer

@typesafeai Can't read + can't write + 20 words for $0.50. Peak satire.

@typesafeai To be fair, this is like using a hammer on a doorknob and critiquing the utility of doors.