Live data from Hacker News

Show HN: Build a website in 30 seconds using GPT-3

durable.co

1–10 of 22 posts

Re: Show HN: Build a website in 30 seconds using GPT-3

#2
Just tried it out I am quite impressed with the demo. I tried a non-given buisness prompt "massage parlor" and it did generate results. Although the provided pictures were a bit strange and the initial description of the business said it closed in 2019. But at the footer of the page it said it started in 2020.

I think this is definitely a step up from "lorem ipsum" default text. The generated text was rather formulaic though. For the "services" section every paragraph started with "is a type of massage that".

Re: Show HN: Build a website in 30 seconds using GPT-3

#3
post #2

Just tried it out I am quite impressed with the demo. I tried a non-given buisness prompt "massage parlor" and it did generate results. Although the provided pictures were a bit strange and the initial description of the business said it closed in 2019. But at the footer of the page it said it started in 2020. I think this is definitely a step up from "lorem ipsum" default text. The generated text was rather formulai…

Thanks for giving it a try! Definitely a lot of polishing to do.

Sometimes the results get super weird (massage parlor is a bit of a tough one), but you can regenerate as many times as you like.

Re: Show HN: Build a website in 30 seconds using GPT-3

#8
post #5
post #4

How well does this website work with open source GPT-3 alternatives? Did you give those a try?

We haven't tried any yet, but would love to. Stable Diffusion for images would be super cool. Which would you recommend?

I was curious about how GPT-NeoX would match up to GPT-3. It seems like the most capable of the freely available models (at least based on their specs). Of course more parameters != quality of their dataset, so I can't really conclude much in this scenario.

As far as I know, https://textsynth.com is the cheapest host and NeoX-20B model is cheaper than OpenAI's Curie. They have a Playground which allows you to experiment without an account.

Re Stable Diffusion:

Based on the results I've seen in https://news.ycombinator.com/item?id=33038117 and other services, if you just want some eye candy I think Stable Diffusion would work fine (and is much better looking than anything DALL-E could generate).

However I wasn't impressed by SD's ability to understand the meaning of my sentence, it seems to just create an image by mashing words together (and sometimes even ignores key words/meanings). According to commenters, DALL-E is much better at "understanding" sentences so if that matters for this site; then I guess you should avoid SD.

Post reply on HN