Comment by deadcoder0904 | Hacker News Reader

Comment by deadcoder0904 | Hacker News Reader

Build something around GPT-3. It's still nascent. Only a handful are profitable.

The top one is making $100k MRR in just 7 months → https://twitter.com/deadcoder0904/status/1398141109075730436...

One sold for ~$1m when it had $20k MRR in 8 months of launch → https://twitter.com/levelsio/status/1377095814963589125?lang...

Another smaller one sold for $4k with $100 MRR → https://twitter.com/heyyfernanda/status/1398679798281605120?...

There are many others getting funded because of GPT-3 → https://twitter.com/kearonis/status/1378832191413702663?lang...

So I'd suggest building it around GPT-3 as it's just a bunch of API calls. Make it profitable & sell it if it's a life-changing amount. The hype around GPT-3 is at an all-time high so big exits should be easy.

And if you want ideas, just copy one of the popular apps & offer it with a discount or make it 10x better → https://www.youtube.com/watch?v=opkHJLVAM4A

Read the Paul Graham essay on ideas → http://paulgraham.com/startupideas.html

You need to solve a real problem so you get paid by people. I have been at it for ~4 years with nothing to show for it so my suggestion is to fail fast.

I have hopefully figured out now how to make a SaaS successful so I should be good for now but you don't waste time. Start now. Fail fast. Use SaaS templates like Bedrock & Style it using Tailwind UI so you can focus on the core logic.

Diving into GPT-3 right now, but they are on beta now, Applied for the waitlist. Meanwhile Im digging into their docs. Thanks for the advice

No problem. It's the best thing you can do to get profitable ASAP but remember to make something useful. Just sticking GPT-3 to anything won't work either. It needs to be something meaningful.

How do you get customers? Also how much do you spend on advertising to get the customers?

That's the hard part & the most tricky as well. As when the product you are making is so simple to make, then marketing is the only differentiation. On a related note, read this blog post → https://www.allencheng.com/starting-a-business-around-gpt-3-...

You need to make something that hasn't been done before so you'll have interest from people as GPT-3 is hyped up now. Spend less than the users will pay you. At first, you need to concentrate on getting customers. Most people use AppSumo lifetime deals but I think they're only good for funding your startup initially without VC money.

Otherwise you can go the VC route & get funded if your idea has legs.

Distribution/marketing is the hardest part in GPT-3 based biz but if you do it right, you'll print millions for sure.

You can even use https://ideasai.net/ to get ideas on your email!

Frankly, those idea bots are useless. You have to spend so much time on them to see the good ideas vs the bad ones.

Even some people who manually tell others ideas are useless as they don't say anything about execution plus they haven't spent a reasonable amount of time on thinking how it could all go wrong. They just spend 5 mins thinking & put it in their idea list.

How does one use GPT-3?

use the openai.com api?

Yes. Apply for access & if they feel generous, they'll give it to you :)

Reply on news.ycombinator.com