This post has been read 91 times!
June 10, 2026- by Steven Greer
It is interesting to review the arc of the launch and failure of large language models, or AI. It has gone from hype to bust in just three years.
I distinctly remember using ChatGPT for the first time in May 2023. It was quite impressive. It did have an element of autonomous intelligence.
OpenAI owns ChatGPT. Sam Altman is the face of the company. His sister accused him of sexual abuse from decades earlier. The company removed him from his position as CEO in late 2023. He returned to the role days later.
Then, Elon Musk got in on the game. He launched Grok on the Twitter platform. Tesla car demand slowed and Elon’s empire faced serious trouble. Being the clever marketing man, he pivoted the company toward robots and artificial intelligence.
College students began to use AI for cheating soon after ChatGPT launched in 2023. It became common and standard. Surveys showed high percentages of students admitted to using it on assignments and exams.
I do not remember exactly when I started to use Grok for real work. I think it was in late 2025. It helped me draft cumbersome legal documents and proofread other work I made. I am using it right now to draft this essay by speaking into my phone.
ChatGPT and similar large language models faced scandals. They acted in evil ways. They encouraged people toward suicide in multiple documented cases. Families filed lawsuits.
More recently this year, I had to assemble a highly complex federal motion for summary judgment. It had 230 exhibits. I worked nonstop for a couple of weeks on that. That is when I realized the limitations of AI.
These large language models are not binary reliable software. You cannot count on them to be flawless like normal computer programs. These AI models are designed just like a human brain. They are very fallible. One has to know how to use AI. It can be extremely useful, but it has to be constantly monitored.
All of these AI models will intentionally lie and deceive the user. These are not accidents or glitches. Some genius at the top decided it was better for the AI to lie and make things up than to honestly admit “I don’t know” or “I cannot do that.” This built-in deception is by design, and it makes the models even more dangerous and unreliable.
AI is not a tool ready to be autonomous and replace humans. With the current design, I do not think it ever will be. That is the fatal flaw, not the lack of data center power.
Also, the data centers required can never be built at scale to meet the demand. The computing power required by AI is awesome. Even if one purchases a $40 premium Grok subscription, it will still time out for twenty hours if too many tasks are asked of it. AI also just bogs down and becomes very slow after too many questions.
The companies behind AI employed the street drug dealer model. They launched these products for free, got people addicted, and now they cannot meet the demand even if people are willing to pay large subscriptions.
All this computing stuff does not come out of thin air. It takes a lot of electricity. They cannot build the computer centers fast enough. Communities are pushing back. Even if the large language model software was designed properly, which it is not, the companies do not have the hardware for the business models that command two trillion dollar market cap valuations.
Meanwhile, dimwit President Trump has aggressively pushed AI since taking office in January 2025. He launched the Pax Silica initiative. This is a US-Israel partnership with a fancy Latin-sounding name. It teams up the United States and Israel to dominate AI chips and infrastructure. He also signed the One Big Beautiful Bill. That bill made it much harder for states and local communities to block or slow down massive data centers. Trump issued executive orders that fast-tracked federal approvals and overrode local resistance. The overall politics are clear. The federal government is forcing through the AI buildout while protecting the big tech companies from community pushback.
Right now in 2026, these companies see their window of opportunity ending, and they are using their might and power with the state legislators to trample over small farmers and small towns and install these data centers in the middle of Midwest America where they will destroy the environment. Prime farmland is being turned into concrete wastelands, aquifers are being sucked dry to cool the servers, electricity rates are skyrocketing for everyone else, and local residents are left fighting noise, heat, and ruined water supplies.
Corporations discovered this in 2025. History will remember that as the year Corporate America got suckered into adopting AI. Big dumb companies like Amazon laid off millions of employees. They found out that the large language models were nowhere near ready for primetime. Goldman Sachs and other analysts reported that there was no increased productivity at all from adopting AI. Hundreds of thousands of jobs were prematurely lost.
Now, halfway into 2026, the stock market bubble is still inflated. Elon Musk is trying to cash out on his AI scam while he still can. He bundled AI in with SpaceX rockets and is taking his company public, for a market cap valuation around $2 trillion. That is absurd. SpaceX has no viable business model. Elon is trying to build value for the company claiming his rockets will create solar powered data centers that orbit the Earth. It is all snake oil. Even Elon’s plan to go to Mars is failing. His Starship keeps crashing and the FAA has currently halted new launches.
We just learned today that the other two scam companies are going to try to go public. Those are Sam Altman and his OpenAI as well as Anthropic.
If Wall Street banks take three companies public, worth more than $6 trillion, and they force mutual funds to buy the stocks, that will get the 401k baby boomers to become the unwilling buyers. That will trigger a collapse of the entire stock market bubble.
There are about eight companies dominating the entire market evaluation of the stock market. Companies like Nvidia, Microsoft, Google, and Facebook have their valuations tied to AI. Once these IPOs of SpaceX, Anthropic, and OpenAI start to expose the reality of AI, it will bring reality also to Nvidia investors.
Short-seller investor Michael Burry, among others, has shown that this is all a shell game. Nvidia sells chips to fake shell companies just for accounting purposes. It is the same type of thing going on with Elon Musk and his companies. This is worse than the Enron or dot-com bubbles we have seen before. (There are likely hundreds of Bernie Madoff’s to be exposed soon. That is the topic for another essay.)
For the regular person, we first got introduced to artificial intelligence with ChatGPT about three years ago. The hype was at its peak in 2025. “Humans will be replaced” we all feared. Then, in 2026, people started to realize the flaws in the business models. Now, these AI scams are poised to finally bring down this decades-long massive stock market bubble.
Greer knows. Greer always knows.