AI left us with only 3 things that matter

Hi, it's Toha. 

I have a month to build a startup. But this issue is not about it.

It is about an idea I believe is really important. Thinking about it for some time probably will change my life, if not changed yet.

This idea is why I believe I can actually make this startup.

I will say "I believe..." many times in this article. That's because I'm not sure I'm correct. I will not pretend I am, that would be lying to you.

If you thing I'm wrong about something here -- let's debate, I'd love that!

So you probably heard about this trendy thing called AI, right? Everyone and their mother tries to predict how it's going to affect their job, their future, and humanity as a whole. I will not pretend I know the future, or even the present really well, but I will give you a very broad perspective, or, rather, a framework through which I see AI and I believe is useful. It's so general you may call it obvious, and it's great if so.

Although AI is a fundamentally new thing in many ways, in one way it's just a continuation of a trend: it's a tool that makes completing some tasks easier. Pretty much any new tool or technology does that, so nothing really new here. 

Let's imagine there is a scale of how difficult some tasks are:

There are three main threshold of difficulty: 

1. Practically impossible. We may consider this on different scales: for one person, for a team, for a huge company. But basically it's a task that even if physically feasible, nobody would reasonably expect to be completed. 

2. LFTWH, meaning Long-Full-Time-Work-Hard. The name says it all -- to complete the task you have to spend a lot of time building skills, researching, preparing, working, and so on. Think of something you complete at work over a month or more.

3. Trivial. This means it's barely a task, something that once set up, can be fully automated. You don't have to sum up the numbers in a spreadsheet -- it's done automatically once you created the right table.

What happens when a new tool or technology appears? 

Some of the tasks slide somewhat down the scale. That's why we have tools, right? They make things easier. 

And I believe what we witness with AI today is a huge Global Difficulty Fall. A lot of tasks become easier. Some of them significantly. I think everyone will agree with this.

But some of these local difficulty falls inside GDF are more interesting that others. Especially those where tasks cross one of the thresholds above. 

Some things which were practically impossible are now just really hard.

That's what world's smartest minds in AI labs and other labs using AI do. We, general public, are not there, so we won't stop here for long.

Some things which were LFTWH are now relatively easy. 

Some areas where you used to need several years of studying and practice to accomplish now can be done in several months. Projects which required three months of dedicated work to complete now can be finished in a week.

I believe this is somewhat controversial, and many will say "No, AI is dumb and it will never do work as well as a human does", or "You used to spend a week writing a code but now you have to spend a week debugging what AI coded".

First, why this is not correct.

AI itself won't do in a minute what a team of people require a month to do, right. But that's not what I'm saying. What I'm saying is that a team of people who could do the work in a week, can now do it much faster using AI. People are still there. AI is still a tool.

Second, why we hear people say things like this a lot. I think part of the reason is they are just afraid of uncertainty. And that's fine. We kinda should be. But simply denying the problem will not solve it. Adapting will.

AI is not superhuman(at least for now), but humans using AI are.

Some things which were at least somewhat difficult are now trivial.

At some moment in the past accountant had to calculate sums of many numbers by hand, with pen and paper. It was really an above LFTWH task. Then calculators appeared which made this job significantly easier, moving a big part of accounting below LFWTH line. Then computers and spreadsheets appeared and got it all the way below trivial threshold.

Of course, there are still accountants, but one accountant can now do in an hour what a team would have needed a month to do before. 

Things which moved into the trivial region with AI are draft writing, simple image creation, summarizing of any kind, transcription, and so on.

Ok, look, I get all this "AI drawings are boring and I will never choose it over something created by human" stuff, I really do. And I agree with it, when we are talking about something emotionally important, like putting a painting on a wall in your bedroom, or maybe even a wallpaper on your phone.

But when I need a simple background pattern for a website, or a sprite for a game that will be seen for 3 seconds, I don't care that it's made without a human soul. 

When I need an icon of a cup, I will choose the easiest and fastest way to get one: opening ChatGPT and asking for one. 

The bottom line, more and more things become trivial.

I believe when new technology makes tasks cross any of these thresholds -- we get most noticeable impact on our lives. It's when some jobs disappear, and new ones replace them. When projects that required a team become possible to be done by yourself.

As a result, if you notice these threshold-crossing tasks and learn to complete them, you have a huge advantage over those who lag behind and do it ‘the old way‘.

And that’s why AI is a big deal, because it’s a GDF. There are many tasks which crossed thresholds and many which will soon. We are still early to grab the low-hanging fruit.

In one sentence: it’s never been so easy to make cool things.

So please, make cool things.

And that’s exactly what I plan to do this month, October 2024.

I’ll be building a startup, because it’s never been so easy to do it. Just two years ago it’d take me months to learn all I need to do it, but now I believe I can do it in a month. I deliberately don’t tell details now, but encourage you to follow my progress and see how it goes.

The idea is not huge but I hope it will be a nice small cozy project.

I’ll be posting all relating stuff on Twitter: https://x.com/imTheToha

However it ends, I will write another issue at the end of this month with a retrospective.

Thanks for reading, hear you soon,

Toha.