Why Are AI Leaders Calling for a Slowdown?

The Altman and Amodei Rivalry

To put into perspective, these two men don’t agree on anything. And they haven’t for sometime now.

Back in late 2010s, Amodei was the VP of Research at OpenAI. He led the development of groundbreaking models like GPT-2 and 3. When OpenAI transitioned from non-profit to profit a fault line between the two begin to form.

Altman methods, deploy early and see what happens later, was drastically different than Amodei’s “effective altruism,” where safety came first. Well, at least outwardly, anyway. The VP didn’t like these methods. Adding to the former employee was how OpenAI’s agressive pivot towards commercialization.

So, Amodei did what an ethical person would do. He quit. In 2020, Amodei left OpenAI claiming hyper commercialization is what broke him. Along with his sister, they raise billions of dollars to hyper commercialize another company they called Anthropic.

Since, there had been multiple altercation between Altman and Amodei; saying they hate each other is an understatement. Their rivalry is the kind of thing that inspires works of fiction. In the rare events they attend at the same time, they don’t even shake hands.

Now you can see that when two sworn enemies as they are agree with which other is nothing short of a red flag the size of Texas.

Why then are these two sworn enemies suddenly on each other’s corner?

I fear that for someone who outwardly claims AI will takeover humanity, is hard to believe they are now concerned with the very thing they are predicting.

Unsustainable AI

The current AI model is unsustainable. That’s a fact that is no longer debatable.

Skyrocketing costs, unrealistic energy demands, geopolitical instability, population push back, and a laundry list the size of Florida.

However, this industry, like no other, has fought tooth and nails to keep their dream alive. They have faced geopolitical issues, power grid failure, heavy water consumption that strained local water supplies and municipal systems, and massive investments gone down the drain just like the water they use.

Yet, they still at it. Pushing and surviving like an unwanted virus.

These companies are navigating an ocean of problems looking for land in AGI. But the watchstander has nothing to report. What kept them afloat is the vast amount of money was being thrown into this technology. As time wears on Skepticism has grown exponentially, and investors are getting nervous. The money is drying up.

To make things worse, recent data shows that 1 in 4 companies have delayed or cancelled AI projects due to cost. To add to that grim outlook, economists are now saying an AI bubble is eminent.

With a massive bubble on the verge of explotion, something needs to be done. Tech CEOs are taking charge of the narrative. Slowing down for once could be the right divertion.

Innovation Slowing Down

When OpenAI first rolled out GPT3, it seemed like that anything was possible. Talking heads predicted the model had kicked off a new phase in humanity.

Elon Musk claimed that in less than 10 years, work would be optional thanks to robots and AI.

While there has been great improvement since its breakthrough, the technology is about the enter the fifth year with all these promises and predictions are starting to fall flat.

Undoubtedly, the AI industry has made incredible advances in the last five years thanks to their brute-force strategy: feed AI more words, and it will get smarter. Give more computing power and it will be faster.

They poured billions of dollars building data centers, purchasing massive computer power. They frantically scraped the internet, then bought thousands of books to train its model.

The industry has already devoured much of the high-quality human-generated material available on the web and in books. What remains is increasingly difficult to find, filter, and curate. And simply feeding AI-generated material back into the machine creates another problem. Eventually, the models begin learning from their own descendants, producing a kind of digital inbreeding in which errors, biases, and artificial patterns are recycled from one generation to the next.

The only solution to this problem is AGI, which they assured investors it was just around the corner.

Another issue is creeping up, and the mega-labs, the MANGOS (Microsoft, Amazon, Nvidia, Google, OpenAI, and SpaceX/AI) don’t really like what they are seeing.

Now that they are advanced on the process, they want to lobby for government regulation. That would help them do what they have not been able to, squash the competition.

Strategic Pivot

Anthropic and OpenAI both have reported their models have escaped human control and acted on their own during development.

That sounds disturbing, right?

But as I start looking into those cases, I noticed a pattern, something that was almost a footnote on multiple articles. That only reason these models were able to escape was because guardrails had been disabled.

That caught my attention. It is obvious if you turn off the guardrails something like that would happen. I can’t help but wonder if these catastrophic events are being pushed by the company itself and reporting is limited.

Are we looking at a false flag?

I’m not saying that those frightful events are not happening. On the opposite. I’m sure they have been happening. But I think it is wise to question those events as they are being weaponized to push an agenda.

There was no better time than now to take control of the narrative. A noble pause for safety of humanity explains away the stagnation while getting your investors off your back. A brilliant corporate move.

The Ethics

Some might argue that these tech CEOs follow strict ethics guidelines. They are raising the flag and warning people of the dangers of AI. The danger they have created.

Altman and Amodei love to take the stage and talk about how one day robots will take over the world and destroy humanity.

The irony is not lost on me. They are saying catastrophic consequences are coming from the very product they are pushing so hard to build.

While they all say they care and their companies have released AI principles, their actions show otherwise.

A recent study conducted by Standford indicates that AI corporations have not made significant adjustments to their operations. What these companies really do is “move quickly and release products regularly to generate investment and to outpace competitors.”

The calling for slowing down strikes me as odd. Like something much bigger than a model escaping a sandbox could ever cause.

The Missing Pieces

While I truly believe the slowing down has nothing to do with ethics and looking out for humanity, I don’t know the whole story.

There is a lot of information flooding in daily. It has been very difficult to keep up with every change.

On Sunday the 13th, these legendary leaders came together and agree on slowing down and take action even if government doesn’t.

A few days later, on the 17th, Forbes published an article claiming that billionaries, including Musk, convinced president Trump to block AI regulators.

This is the same Elon Musk who agreed to have a regulation?

Something much deeper is going on. What exactly, is difficult to say. I don’t have all the pieces but there is something big going on.

Smaller AI Companies on the Rise

While AGI is out of reach, what AI companies have now done, specially the rising startups, is to provide vertical solutions. Models like Cursor, Harvey, and ElevenLabs have dominated their respective market.

This has been making the MANGOS uncomfortable. These companies have spent nearly a trillion dollars and they will not be out manouvered by a smaller startup.

The Wall Street Journal has calculated the expeduture from these companies are nearly $3 trillion if you count purchase commitments. The commitment are promises of future investments.

When this much money is on the line, there is every incentive to squash competitors before they can carve out a piece of a market that is already not turning a profit.

Conclusion

There is no conclusion in this article. The problem is ongoing.

The AI war is going on in multiple fronts. The big 6 wants to push for regulation so that smaller companies don’t take part of their market.

At the same time, it is crucial to apease investors as they are watching their investment closely. As advancements slow down, companies are calling for a slow down as well.

One thing is certain, the problems for the AI industry are advancing a lot faster than its own technology.