If you think about it, there are no AI “agents”, no “swarms”, nothing “agentic” or “identic”. These are just the latest buzzwords for the same invention: the LLM chatbot.
Still, there is a lot of talk about AI replacing certain job categories, especially programmers.
Actually, there are only about 30M programmers worldwide (according to ChatGPT) or about 10M (according to Grok). That’s 0.004% of the population (per the ChatGPT figure). No real jobs are threatened, even though the AI guys say so. Not even the programmers’ jobs are really in danger… as software needs programmers to check and fix errors, and to deal with the real world.
No jobs are actually in danger
So far, the only *confirmed* job I’ve seen being replaced is teenage photo modeling. The teenage model pictures can now be created with generative “AI”. But, that’s not a real job.
We see dancing robots in videos from China, but nobody can even produce a robot that can do the dishes, or anything useful. And if they do in the future; so what? A better dishwasher only saves us time. That’s no revolution.
AI is not advancing in the entertainment industries either. Self-ordering kiosks at fast food restaurants and movie theaters were available before the new crop of AIs (LLMs) came online.
Robot servers are an already existing and boring gimmick in some restaurants. We will probably see more automation, thanks to the fact that computers finally are starting to get what we are saying to them. That is pattern recognition, not intelligence.
Word-guessing affects how people write
At the same time, the word-guessing AIs are having an oversized effect on how some people write, and even think. The producers of AI (LLM) output are the ones most likely to pick up an AI psychosis, thinking they can solve others’ problems in fields they are not experts in.
“This is not X. It’s Y.” is used in almost all AI texts, and now consumed so often that people are starting to imitate it, unconsciously. New trends pop up, like “load-bearing”. All the chatbots use the “new cool” word simultaneously in all their texts. The AI guys can’t even see it.
Experienced writers know to stay away from such clichés, but regular Joes don’t seem to mind the “empty” sentences when creating texts. Those texts are not impactful. Non-writers don’t know the difference, though.
They think “what others do can be done by AI instead”. Not what YOU do of course, only what everybody else is doing.
If a sentence can be lifted out of one text and put into almost any other text, it’s useless. It’s just hype.
“Most people are not even aware of this yet” is true of almost any news or information.
It’s typical AI filler “word salad”. Specific words like “delve” and “poised” and “entering a new era” have been popularized by AI slop.
This type of hype writing wears off and people get bored of reading the recurring superlatives. They move on. Remember NFTs? Metaverse?
The LLM-type “AI” is not the end-all be-all. Yann LeCun is right.
The chatbots are not even replacing the 30 million programmers. It’s just another tool. Yes, it’ll enhance productivity here and there somewhat. IBM now says 6-7%, and they are hiring interns again, after discovering how bad AI results are in real life. Klarna is rehiring support staff. “AI” messes things up too much.
We’ll manage and move on to the next hype. Musk and Altman will come up with some new scammy hype and the scrollers will believe it. They never read a book, or have even really understood what technology is. They think it can be dictated, when rather it’s discovered.
Why won’t the LLM “AI” reach the goal of a functioning AGI? Because you need intent for programming to be interesting. Your chatbot has no intentions, no will, no heart. I may be delusional and tell you that it’s conscious, but it has NO AGENCY, NO INTENTION.
It’s like an advanced calculator, but for words. If you think it’s alive, you have AI psychosis.
So instead of talking about the future, let’s look at the here and now. What was promised by the AI salespeople, let’s say, two years ago?
Here is a rundown: Most of the so-called “AI experts” said that around 50% of all technical jobs would be gone by the end of 2025.
WHAT? Do you remember?
They all said that at least 30-50% of all jobs would be replaced by AI by now.
As you well know, as you see around you, that is not true.
So, why are the self-styled so-called “AI experts” still invited to tech podcasts? Their predictions are completely wrong, colored by their own psychosis.
They promised “agents”.
Wow, what a cool word. Like something from the Matrix.
The geeks were flabbergasted and appropriated it, thinking they are now running “agents”.
It’s stupid, but the fanboys repeat it. Like there were any “agents”. It’s still just the LLM, the “chatbot”, run in several instances and with prompts like “make no mistakes”. How can people fall for this?
And “they” can hack into systems.
With “they” I mean the scripts started by ordinary humans. “They” are not “they”. It’s an “it”.
These are just programs running. There are no AI:s. Just programs.
But the word “chatbot” isn’t cool anymore. Watch out for the next buzzword. Dummies are now vlogging about their “11 agents working 24/7”. Yeah, right.
It’s still just the same thing as before. The LLM chatbot. Nobody has demonstrated they can make any money using “AI”. Because there is no AI.
The whole thing is very similar to the NFT phase in crypto. Every other month someone had to come up with a new buzzword to prolong the trend. The same thing is happening in “AI”. Just like with “blockchain”, if you remember?
And then we have the problem of motivation. Human intelligence is often driven by a want or need. The so called AIs don’t “want” anything.
Last year, in 2025, Sora was introduced and it could produce about 6 seconds of video. Now, in 2026 Seedance is promising a jump to 15 seconds. If we extrapolate this, it would mean: 30 seconds in 2027, right?
In 2028: 1 min, in 2029: 2 min, in 2030: 4 minutes… 8, 16, 64 and in the year 2035 we finally reach the stage of full-length 128 min movie creation.
Let’s say they can build datacenters even faster and we get there earlier, at 2030?
So what?
The world is not going to change because people get some new movie-making tools.
On the contrary, we are going to need new tools to sift through all the crap. Amazon Books, the original Amazon, is already almost destroyed by “AI-written” books. Jeff Bezos has had to introduce a new rule: max three books can be published per day per author.
This AI spam slop will show up everywhere where algorithms are automated. And they will destroy value. The AI psychosis has mainly struck AI company bosses.
So, we are in a transition phase. New tools will have to be developed to help people navigate the AI slop. As Amazon, Instagram and others don’t seem to care about its future, letting AI slop destroy the output; the field is ripe for innovation. “AI” will dazzle us again, in some new form. And create more jobs and industries than it destroys, IMHO. But, it will take a long time, during which new buzzwords continuously will have to be created by the hyperscalers, to keep the interest and investments flowing.
Already we are seeing lots of great media being created that never would have been possible before the advent of image and video generators. Specialisation in programming, helping people and businesses create their own apps, should grow immensely going forward with new programming tools speeding up production of software. New math software should help speed up scientific discovery as well. New applications will pop up. The future is bright. And jobs were never in danger.