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 :)

Time in the market always beats timing the market.

I like that!

Doge peaked but it can still go up when there's Elon Musk behind it. That man can move markets with his 55 million Twitter followers.

There's all kinds of coins blowing up. These are mostly shitcoins but they do provide 100x returns for some.

I recently read a thread when someone turned their $2k investment into $1m in 1 month by investing in SafeMoon.

Coins like this are popping up everyday. There's also a burst of revenge coins like https://www.fckyousarah.com/ when someone's GF cheated on him so he made a revenge coin dedicated to her. And people are loving it as it's really stupid & hilarious.

Ha, I recently asked one such question & did not expect to see this so soon: "Ask HN: How do you spot trends in Crypto before it blows up?" → https://news.ycombinator.com/item?id=27219070

How does this work in theory?

A coincidence! Hadn't seen that, but wish I had.

I run a cron job every minute that pulls in the latest trades from the Binance Smart Chain, and calculate the price of the token on every trade. If the coin posts two consecutive hours of > 10% price rise (and meets a certain volatility threshold), then I consider it trending.

Once it falls below certain price/volume thresholds, I assume it's passed it's best and it moves into 'recently trending'.

Where did you get this data from?

I filled it myself using Google to find projects I wasn’t aware of.

This is so fucking cool. `cat` doesn't work in Terminal. Probably put a bunch of basic commands for someone curious.

`cat` works just fine, but `more` or `less` are "not installed" :-)

I did `cat about` I think which was the wrong thing to do. Didn't know that was a folder. It works fine.

You probably cat-ted a directory, like the "about" directory.

Yep, I did `cat about` I think which was the wrong thing to do. Didn't know that was a folder. It works fine.

That's strange, `cat` is supposed to work well. What did you `cat`?

I did `cat about` I think which was the wrong thing to do. Didn't know that was a folder. It works fine.

I used `cat my-dream.cpp` with no issue

I did `cat about` I think which was the wrong thing to do. Didn't know that was a folder. It works fine.

Worked for me, what did you `cat`?

I did `cat about` I think which was the wrong thing to do. Didn't know that was a folder. It works fine.

> One highly successful software entrepreneur is burning through something like $80 million in venture funding in his latest startup and has practically nothing to show for it except for some great press clippings and a couple of beta customers — because there is virtually no market for what he is building.

Who was he talking about? Anyone has any clues?

The number of possible startups this could be is... incredible.

Remember “color” for example? This happens more than you’d think.

Wild guess, Magic Leap?

Augmented reality has a potential market as large as that of cell phones. Magic Leap raised not tens of millions but more than a billion dollars if I recall correctly.

I believe the issue is that their product is expensive and therefore did not become popular and therefore has relatively few applications and so does not become popular.

Wearable augmented reality will probably be very big some day. (The probably being mostly a hedge against essentially the Glasshole factor--i.e. never becoming socially acceptable for day to day casual use, as opposed to commercial applications.)

But the when is anyone's guess and could easily be a decade or three out.

"Easily three decades" is a short-sighted prediction in my opinion.

Some of the optical waveguide prototypes are quite similar to normal glasses especially with power and compute on a tether in a pocket or something. I will be very surprised if it really takes ten years.

It's mostly the observation that VR and AR have been right around the corner for a long time. (And, yes, I'm aware that are VR products in the market but also that it's not really mainstream.)

So I wouldn't be shocked if there were mainstream consumer AR glasses before the end of the decade. I also wouldn't be shocked if 10 years from now they were still at the "real soon now" stage.

Thank you, your comment made me check the summary of it & I found a good one on Farnam Street → https://fs.blog/how-to-read-a-book/

Now I gotta give that book a shot :)

There is GPT-Neo although it doesn't have as many parameters as GPT-3. You can also check out this awesome resource → https://www.reddit.com/r/GPT3/comments/mj5nn0/gpl_prompts_re...

Also see → https://news.ycombinator.com/item?id=26696735

Well, what's the issue?