How a med student made the best AI writing assistant and got acquired by Jasper | Hacker News Reader

How a med student made the best AI writing assistant and got acquired by Jasper

theexitgame.substack.com

67 pointsmujavid21 comments

This is impressive, but the tactics he used along the way range from meh (eBay arbitrage) to pretty sleazy (posting fake positive reviews, using bots to spam comment sections).

At least they pivoted into a career space where people aren't surprised by unethical behavior, rather than become a doctor.

I wonder why he doesn't use the GPT-3 to run a content farm. He sounds like the guy for it.

From the article:

> Qasim's guerilla marketing tactics kicked in. He wrote hilarious reviews about the game and posted them on different subreddits. One post went viral and his downloads skyrocketed to 10,000.

This article itself is also likely an extension of the author's self-promotion tactics. It's not uncommon for first-time startup sellers to do a self-promotion tour through blogs, podcasts, and other outlets where they try to shape their online reputation. This is basically table stakes for someone getting ready to fundraise a new startup on their past reputation.

I worked for one startup that basically fizzled out and was sold off at fire sale prices to a competitor. That didn't stop the founder from building his entire online persona around successfully building and selling a startup. Take these articles with a grain of salt.

> This article itself is also likely an extension of the author's self-promotion tactics.

• This is Mujavid Bukhari's only article on a Substack account established yesterday.

• Mujavid Bukhari's Twitter account has tweeted only 23 times since 2020, and just twice this year — starting yesterday, with both tweets being about Qasim Munye.

So yes, this is a purely promotional article placed in an attempt to attract coverage from influencers/media with a meaningful amount of followers.

Are you implying someone who would write fake reviews for his own products under fake names might also write fake blog posts about himself under fake names?!

I for one, believe it was entirely written by a 100%, genuine friend of Qasim. A guy named GPT-3.

> This is Mujavid Bukhari's only article on a Substack account established yesterday.

Good catch. Out of curiosity I looked up his LinkedIn profile. He puts "previously at Google" in his headline, but when you look at the dates he was only employed by Google for a few months.

I think we're seeing an obvious case of "reputation hacking" with both of these people.

As a an engineer, I find it hard to justify creating a product without a "zero to one" or "order of magnitude better than competition" justification.

The fact that Qasim basically just re-packaged the GPT3 API into a web interface and some Stripe billing, and was able to make money is surprising to me, and is basically something I would never consider doing.

It's the kind of project I would almost be embarrassed of telling other people about. But that's the wrong attitude, because, he made money and exited, right?

I suppose the value add Qasim truly offered was in the user acquisition and marketing.

I'm curious how much the acquisition was for. Would companies really pay > $5m for a GPT3 API frontend? Maybe? I've seen dumber decisions made.

I don’t think they paid for “GPT3 frontend”. I think they wanted to consolidate and get all of the users. I imagine tech was somewhat of an afterthought.

This article sounds exactly like that Ryan Holiday book of "Trust Me I'm Lying".

I tend to forget how people become successful using media manipulation.

What's worrisome is how people will use GPT-3 to further manipulate media.

Never really thought about how much you could make from a startup if you sell out before raising any money.

Good move on selling it tbh. Don't really see how much of an edge companies like this have considering it's built entirely on the GPT-3 API.

Congrats to Qasim.

It seems like it's really just reselling a commodity. Which can be fine, many companies are successful doing it, and you could (in theory) build a more useful product on top of it.

There's still the fact that many of these companies are not doing that, and are entirely dependent on GPT-3 continuing to be available to them at the given price point.

It doesn't sound like he made a lot of money doing all this. It's good hustle for an outsider put probably pays less than regular job for someone with his talent.

It sounds like he had 3000 actively paying business users at the start of this negotiation, with revenue matching up with a series A level co (so likely significant, but definitely under $10M ARR. If they started with a benchmark valuation of ~$15M, cut that down a bit, and then negotiated up based on continued growth this could have easily clocked in at high 7-low-8 figures.

I definitely wouldn't count that as "not a lot of money".

Having meetings with Brockman and Altman afterwards also point to it not being an insignificant acquisition amount.

Taking your criticism further is this analysis.

https://www.allencheng.com/starting-a-business-around-gpt-3-...

The best though! Not just any!

And I for one trust that this is an entirely accurate assessment, on the basis of this article he just paid for.

Mm-hm.

Incidentally, I just wrote "the best" file downloader that you should totally buy, which is totally not just a zenity call on top of wget.

It says so, right here in this article I'm about to pay for.

I really hate these (probably substack?) modals that ask me to subscribe after the first paragraph. The f*ck I will. Who does that based on three sentences?

Something about the way that was written brought this amusing thought to mind: what if this hype-man blog post about shortly had been created with shortly.

Reply on news.ycombinator.com