π’ Introducing FaerieAI, ChatGPT's little sister π
- Built on OpenAI GPT3.5 APIs
- No annoying delays/limits/refreshing
- Responses render all at once, not slowly letter-by-letter
- Configure your own settings
- Pay as you go with #bitcoin
Try it π https://t.co/uj54DeeNbY
@samcwl If there were a market for 85% complete app, that could do well
That last 15% is the hardest and without having that solid, Iβm not comfortable productizing this like I planned - may revisit later
As it is, seems logical to focus more on developers than end users just yet
@samcwl If there were a market for 85% complete app, that could do well
That last 15% is the hardest and without having that solid, Iβm not comfortable productizing this like I planned - may revisit later
As it is, seems logical to focus more on developers than end users just yet
π’ Introducing FaerieAI, ChatGPT's little sister π
- Built on OpenAI GPT3.5 APIs
- No annoying delays/limits/refreshing
- Responses render all at once, not slowly letter-by-letter
- Configure your own settings
- Pay as you go with #bitcoin
Try it π https://t.co/uj54DeeNbY
I increasingly see GPTβ3/LLM prompts as assembly code, not as human interface. We shouldnβt be writing prompts, but prompt compilers. A template string is not a moat.
We improved #ChatGPT with filesystem access so now it can code up an entire application from scratch based on nothing but you describing it in words.
Early access starts next week for our waitlist. Free for now but we want ur money.
Join the waitlist π https://t.co/9vW0PThCgq
This is the craziest tweet on this site.
OpenAI engineer calculates the cost to build a Google killer and comes to $50m for the embedding.
$50m for a ChatGPT that knows everything thatβs on the internet, almost instantly.
Unbelievable.
Working well so far: a bunch of people submitted great ideas!
On to the next step: converting these convos into detailed specs
If you had a good convo with the bot and gave it a real idea (we'll be asking it to rate your seriousness :), we'll email you soon w/ spec & next steps
Working well so far: a bunch of people submitted great ideas!
On to the next step: converting these convos into detailed specs
If you had a good convo with the bot and gave it a real idea (we'll be asking it to rate your seriousness :), we'll email you soon w/ spec & next steps
@GlavinW Cool, thank you for the feedback - it has a limited memory cuz it only 'remembers' the last X lines, whatever we can fit in the prompt
Will improve its memory over time
But if you answered the questions once, that's enough for our purposes
Thank you!
@GlavinW Cool, thank you for the feedback - it has a limited memory cuz it only 'remembers' the last X lines, whatever we can fit in the prompt
Will improve its memory over time
But if you answered the questions once, that's enough for our purposes
Thank you!
@GlavinW Our first feature request! ππ
No reset yet - For now you can just put the requested info anywhere, it won't need to be exactly sequential cuz all this will be later parsed by more AI that is smart enough to know
Will be tweaking the bot prompts over time so it's less stupid
@GlavinW Our first feature request! ππ
No reset yet - For now you can just put the requested info anywhere, it won't need to be exactly sequential cuz all this will be later parsed by more AI that is smart enough to know
Will be tweaking the bot prompts over time so it's less stupid