AI makes you way less productive unless

When talking about artificial intelligence, there seem to be two main camps out there. The AI bros and the luddites.

AI bros share the Dario Amodei attitude. The great replacement of all the jobs. They grin talking about how AI is going to make everyone obsolete (except for them obviously because they know how to "proompt").

They laugh at people who apparently wasted time learning coding, design or marketing.

I can just one-shot all your years of experience in one prompt, bro! You're done!

On the other hand we have those fighting AI tooth and nail. Their argument often starts with copyright infringement. Yes, AI has trained on content from humans, often without permission.

There seems to be almost nobody "in the middle" of the chart.

They talk a lot about craft, human touch and battle capitalist plans to get rid of humans in the workforce.

These people most often shun AI completely. Don't use it out of principle. I can respect that, but at the same time it feels like the Red vs Blue race all over again.

Another multi-model tool accepts the pricing reality

Truth is in the middle

Reality often doesn't care about hot takes. It comes at us with things that just happen, and very rarely they're one sided. Recent raising of prices and issues of big, frontier model AI companies prove that perfectly.

It was transparent and unpredictable all at the same time. And many of the AI bros online seem betrayed. Complaining, switching allegiance to a different provider, being loud about it.

The expectation was to get an inexpensive cheat code to get ahead in life. The reality caught back with it being just a hook up scheme.

But when this happened, the anti-AI crowd cheered for all the wrong reasons. Predictions started to surface that AI is "completely over". Good riddance!

In both cases it seems like some magical being is being glorified or vilified.

While in reality…

It's an inference search engine

AI is not as much of an intelligence, as it is the next iteration of a search engine. If parsed a lot of information from the internet, and by predicting the next word, it "generates" information you'd normally be searching for.

Wikipedia says: Inference is a logical conclusion reached by combining evidence, observations, and facts with background knowledge or reasoning. It is the process of moving from known information to new, unstated insights. Common synonyms include conclusion, deduction, reasoning, assumption, interpretation, and presumption.

Which means a decade ago, when you searched through Google on how to do something (coding tips for example) you had to infer it yourself. You combined what you found with what you know and used your experience, logic and knowledge to reason for or against it.

This is why it was so seamlessly integrated into Google's very own search results. It's just another layer of what we all know.

See? When explained by AI itself, it tells you exactly WHAT it is.

What AI does, is averaging popular inference patterns for you. One less step, but it's still mostly a search engine and a mix of knowledge from other people as the filtering experience.

Don't get me wrong, it can reach some pretty groundbreaking conclusions. But it can also do the exact opposite. It can lead you astray, while praising your amazing insight in prompting at the same time.

And after most of our lives being proven wrong by society, it does feel good to hear you're absolutely right. Even when we're dead wrong.

Meme found on Reddit, AI's being supportive of bad ideas is by design, to keep you hooked

What is wrong with that?

Here we're getting to the main issue of anti-AI crowd. They claim all of AI is stolen knowledge. The discussion obiously gets loudest around generative part of it all. And I get it. It's a shorter path to plagiarism of some kind for sure. Before, you had to find a book or a post, and steal that or try to paraphrase it.

It is definitely amplified, but all of creation is merging inspiration. Previously it just happened as signals between neurons, now it rides on a neural network.

Stealing is everywhere

I know this firsthand. My books and articles were put into their "original books" by many design influencers from all over the world. One Indian "brand" went all in and basically copied an entire free book I made, down to the color scheme.

People were fine with plagiarism before AI.

The issue here may be that this time it's unconscious plagiarism. As before you had to consciously copy and paste some text to pretend it's yours.

Now you have no idea how it got generated. Was it a mix of a lot of things, or an almost 1:1 copy from a single source?

It's not my code

There's this famous meme with two designers arguing who stole whose design, and then two developers acknowledging neither is really the author of the code.

And for some reason most anti-AI people don't mention completely scraped stack overflow (which later got blocked), reddit and source code of hundreds of products to train the AI models.

As if code is not creative. As if that part of stealing is fine.

But when it gets to artwork or written word, it stops being fine because that is the domain of the true artists.

See how nuanced it is?

See how rocky that relationship was?

Hunter S. Thompson famously said he learned to write by re-typing entire books by F. Scott Fitzgerald and Ernest Hemingway to learn their prose rhythm and "feel what it feels like to write that well".

Then he wrote his own books.

Anti-anti-anti…

I tend to be contrarian in life. Not for the sake of being against people, but because in most cases I simply don't feel like ends of a spectrum fit me.

I also don't like being put in boxes.

I don't vote left or right. I'm not for or against AI. How I try to pick a path is more nuanced. And I try to change my mind often, when someone convinces me I was wrong. Or not fully right.

Using AI

I use AI to help me code. I use it as a search engine sometimes. Sure, you need to double-check the outputs. But you always had to do that. Same with manually found reddit page explaining how to achieve some effect in code.

I can't see myself using generative AI to write articles. Or video scripts. Or make non-meme like images (or generative gradients) of any kind. So I do draw a line somewhere too, but it's not black and white.

I still do the UI, logic, app architecture and animations by hand.

The interaction you see above was coded, then recorded 30+ times and analyzed frame by frame. Then tweaked in the easing, length, movement, bounce and layer order until I was fully satisfied.

Coding with AI doesn't have to be "all vibes and no understanding". It's up to you how you use it.

Frame by frame recording for a single interaction. This is the human touch.

It's a tool

When you dismiss the delusional notion of intelligence, you can see it for what it is. A tool. A way to find (and sometimes arrange) information faster.

You don't have to infringe copyright. You don't have to plagiarize. You don't have to go against your own principles. But getting information faster, when you know how to verify it, is a superpower.

Is it perfect? No. AI outputs, even from Opus 4.7 are still not good enough for 80% of my professional work. I test it out regularly to see what I can actually outsource.

But coding quick prototypes, or looking for patterns across 8 pubmed articles is where I do feel the benefit.

The real problem

The real issue is not in AI itself. The problem is in the multi-level rat-race. On the lowest level we have extreme fomo of being replaced by AI. Everyone employed by someone else feels this. And fears this.

This is where the classic (and mostly false) claim originated:

AI won't replace you. Someone using AI will.

This leads to the often insane push for AI among big companies, when only 27% of those who use it large-scale report any visible ROI. That is level two. The corporate stage.

They seem to can't wait for their own predictions to come true.

Then, there's the contradictory claim from the likes of Anthropic CEO saying that 50% of all white-collar jobs will be replaced. Talk about "someone using AI". Unless he meant himself here.

Then on the other hand we have Sam Altman starting to realize claiming AI will "take jobs" may not be a winning strategy, even if ultimately it is the goal. So he changes his tune to attract more users, potentially scared of AI because of the whole replacement debacle.

Is he just making a PR stunt here, or is this genuine? ;-)

But the actual battle is happening at level three. This is the most ruthless one, among the frontier AI companies. It's a festival of burning investor money, causing more fomo, actual innovation and fight for survival.

The real goal

The real goal is not replacing all humans with AI. That is just a side-effect. What those companies need is to not lose. Yeah, you heard me. It's not about winning anymore. There's too many factors, too many black swans, too much noise, and totally not enough value.

Local models are getting increasingly powerful. Gen Z is outright boycotting AI. The majority of humans still don't feel the need to use ANY ai. Out of those that do, the vast majority uses free AI. Whether it's AI based, or just trials, it's completely enough for them.

And if those free options go away? We're not sure what'll happen. Some may start paying, but the danger is most will just discard it as a gimmick it was (to them) and stop.

You need to need it!

The first step of the goal is to convince as many people as possible to use AI. Or, alternatively use it as a control mechanism and advantage in government, military and strategy.

Accenture purchases copilot licenses from Microsoft for almost 740,000 employees.

Judging by the involvment of the AI companies with the governments, it seems like the B2C model is slowly fading. Even the traditional B2B is taking a hit, and the 728K paid seats for copilot from a single enterprise won't move the needle much. At least as long as those people don't really use it.

The future may just as well by B2G.

And in that case only the richest consumers and companies will be able to afford AI. Either as access to frontier models, or the ability to run good ones locally on powerful hardware.

AI may just as well shift primarily to government use, where the most capital is.

It's just a phase

When you think about it from this angle, it doesn't really matter as much whether you're strongly pro-AI or against it. The final decisions will not be ours anyway.

Where AI ends up is unpredictable, but judging by the current progress, we may have reached the limits of what LLMs can do. Maybe it's time for another breakthrough.

A different direction.

Or stalled, iterative, 1% improvement sold with even more fomo online.

Instead of fighting the other side, it may be better to just figure out a way to use the current generation search engines to your advantage. You can do it while staying morally fine. Nobody's asking you to generate artwork. Or to write books.

And AI bros will likely jump onto a new fad once this becomes unsustainable financially. They already did crypto, then NFT's, then dropshipping. AI is just another way to get attention because everyone is taking about it.

But it's also super tiring. I mean it.

Switching some of my workflows to paper is sooooo refreshing.

Relax.

Take a deep breath.

And I'm telling this to myself too. It's time to cover other things too. Not get pulled so much into that AI-everything world. For a while it seems like it's the only thing that exists.

It's not.

There are interesting non-AI workflows that I use daily. There's inspiration outside of the digital world that seeps into it.

That is my cue. I'll still cover AI, but I choose to not subscribe to the constant fomo of the news cycle. All-AI tech is boring.

I strongly believe the healthiest approach to this game is to use the game pieces from time to time, but not to play like the game masters want you to.

So throw away the instruction manual, and assemble something completely of your own. Whether it's with AI, without it, in the real world or in the digital.

Am I an AI bro? Or a luddite?

I'm neither. And that's my decision.

How about you?