The Hated One

Why you shouldn't believe the AI extinction lie

Added 2024-05-03 19:34:21 +0000 UTC

There is a powerful motivation to keep you thinking that AI is an existential threat. [0]

That we should treat it with the same level of urgency as a nuclear war or a global pandemic. [1]

And yet, we shouldn’t stop developing AI. We should accelerate it as fast as we can. As long as it’s the select few good guys that get to control it. We must not let AI fall into wrong hands. [2]

But there is a growing opposition to this movement. A conflict is arising. There is those that want to keep AI closed off and tightly controlled and those that want to leave it open and accessible to all. Even though both sides claim they are doing for humanity, only of them is right. This is an arms race for who gets to dominate AI development and who will be left out.

I am the Hated One, and I make explainer essays like this one, so far still without any sponsors… or money… or friends… So let me show you how you are being manipulated into handing over all of AI technology to some of the most powerful corporations.

2023 was the year when AI exploded into the mainstream. GPT-4 was just released, with chatbots, generative images, and AI videos flooding the Internet like a hurricane of sensationalism.

But behind closed doors the big tech pooled hundreds of organizations in a massive campaign to lobby the US federal government. The world has never seen such an organized lobbying effort. The number of AI-lobbying orgs spiked from 158 in 2022 to 450 in 2023. Which didn’t just include the usual big tech culprits, but chip makers like AMD, media moguls like Disney, or big pharma like AstraZeneca. In total, they spent $1 billion on lobbying. $1 billion that among other things went to persuading law makers to get AI regulated precisely how they wanted it. [3] [12]

So what did they want? Well, all of that lobbying culminated in a bipartisan bill proposed in the US Senate. The bill would have the federal government regulate artificial intelligence nation-wide. It would create a new authority that any company developing AI would have to register to and seek a license from. License is just a different word for permission. They would have to be monitored and audited by federal agencies and they would be held liable for any harm caused by the use of their AI models. [4] [5] [6] [14]

Which, you’d have to ask yourself – how many new startups would have the funds to comply with such a strict licensing regime? This would mark the end of open source AI, because nobody would want to give anyone open access to their AI model that could hold them liable for abuse. Only big, closed, proprietary models would survive this. [7] [4]

Oh, I am gonna be mentioning open source quite a lot. Open source simply means that anyone could use or distribute software freely without the author’s permission. Yeah, we could have had that, they just chose not to. [8]

How could this legislation proposal even be crafted? Well, it wasn’t by accident. The two US senators proposing the bill had multiple hearings with OpenAI, Microsoft and Anthropic, the biggest players in the industry. Their witness testimonies lead to the drafting of the bill, which was later endorsed by an obscure organization called Future of Life Institute. [5, 6] [10]

You’ve never heard of them before, but you’ve heard of Elon Musk signing a letter calling for a 6-month pause on AI development or the world would end. That was the Future of Life Institute. [0]

Of course, nobody actually paused AI development. Everyone who signed the letter went back to their work developing AI faster than ever before. [11]

But governments should totally “step in and institute a moratorium”. Sam Altman didn’t even sign this letter. [9]

But he signed another one from another obscure org called Center for AI Safety. This one came with a simple statement – “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war”. [1]

Both of these stunts were picked up by media, which served well after years of conditioning that AI poses an existential threat, far greater than anything else. But what’s the implication of treating AI with the same level dread as literal nukes? You gotta prevent proliferation. You can’t allow free and open access. Only a handful players should be allowed development of this technology and they should keep it closed, confidential and proprietary. You can’t allow this technology falls into wrong hands. [7] [13]

Which makes sense, if they are right. But they are not right. First of all, there is actually no proof or consensus that future superintelligence is possible at all. This is the first predicament of an AI apocalypse. But it has no merit. [14b]

Yes, there is a non-zero chance it could happen, just like there is non-zero chance we could get invaded by aliens. [13]

The second premise claims that AI will continue increasing its intelligence indefinitely. This is false. Our current AI models are already reaching a ceiling – training data is running out. Quantity-wise, there is actually little space left for scaling. AI generated data eventually becomes poisonous to the point models deteriorate. Computation is becoming increasingly more expensive both in terms of operational costs and resource costs. AI would need major scientific breakthroughs in order to significantly increase its intelligence from where it is now. [14b] [15]

The conclusion is therefore also false. It’s not certain at all that AI will come close to human intelligence levels, not to mention surpassing them. We shouldn’t act as though AI is a threat greater than a pandemic or climate change.

There is actually a powerful billionaire group that is bankrolling research, YouTube content, and news coverage pushing the idea of AI as an extinction risk. It’s an ideological monolith of the longtermist effective altruism movement, which is a rabbit hole so deep this will be its own video so let me know if you want it.

But to give you an abridged version, billionaire philanthropies are paying fellows and staffers working closely with governments, think tanks and policy groups to influence local and federal legislature and help draft executive orders. They are working to persuade them to focus on future hypothetical threats of AI, while simultaneously trust them that the AI they are developing will be the good guy. Fear AI except when its our AI. Then you should trust us completely. What we have here is a fantastic case of regulatory capture. Nobody from regulation had enough expertise so this organized group of big tech companies and effective altruists stepped in to fill the power vacuum. [4] [12] [16] [17] [18]

But there is a growing number of those that stand strongly against all of this. They say that if we do this, if we allow licensing and strict regulation like the big tech lobbies for, it will be the end of open access to this technology. It will lock almost everyone out of AI development and will leave only the few powerful incumbents in the game with closely guarded proprietary AI models. Open source alternatives that could be distributed for free would be regulated out of existence. [4]

Professor Andrew Ng is someone you don’t see sensational headlines about too often. But he is a key figure. He is the one that taught Sam Altman from OpenAI and he stood behind AI projects of Google, Baidu and Amazon. And now he says that the big tech is fearmongering policy makers into drafting legislation that would kill their competition. He rejects this idea that AI could pose an extinction-level threat and he thinks the big tech is using fear to damage open source AI. Because open source would mean anyone would have open access to this technology. [7] [13] [21]

His not alone in this thought. There is a growing counter-faction of scientists and researchers that are also calling out the big tech’s true motivation. They too argue that this is just an attempt to hijack regulation to cement incumbent AI companies and to focus policies on future existential dangers instead of addressing current and immediate problems. They warn that the licensing regime the big tech is calling for would monopolize AI development as they would be the only ones able to accommodate it. [19] [4]

So what is the solution then? How can we prevent this small group of the most powerful companies in the world to capture AI market for themselves?

There is this secret document written internally by a Google engineer that leaked online. The engineer says that both Google and OpenAI are losing the AI arms race to a third faction. This third faction being open source.

This document is a beautiful read of a terrified mind that realized they’ve been doing AI wrong all along. It details how Google slept at wheel while the open source community got way ahead of the game by focusing on smaller scale models more appropriate for the end user. He lists multiple open source AI projects that do what Google’s or OpenAI’s large models do with a comparable quality but at a lower cost. How these open source projects solved the scaling problem with better quality data and how competing with open source is a losing battle. I love every single word of this letter. And this is where your contribution steps in. [15]

There is an open letter from the Mozilla Foundation, the guys that make Firefox, that calls for opening up the source code and science of artificial intelligence. This letter was signed by Andrew Ng, of course, but also by Jimmy Wales, the founder of Wikipedia, by folks from Creative Commons, the Electronic Frontier Foundation, the Linux Foundation, academia, and even a few souls from the big tech. [20]

This letter got next to zero coverage in the media. But it’s clear this is what the big tech fears and wants to prevent with their lobbying power. They don’t want regulators to realize that open source AI might be better, more equitable and safer in the long term. Open source takes power and control away from the top players and gives it to anyone with a laptop. Open source allows public scrutiny and accountability. It’s what allows researchers, experts, journalists and users to audit, question and verify what’s going on. This is what can earn people’s trust because it allows everyone to participate in making it better rather than just trusting a selection of executives to do what’s best for humanity after they serve their shareholders.

This is where you can play a role. You can sign this letter too. And you can also support open source projects that work on democratizing access to artificial intelligence. Rather than paying for premium subscriptions for proprietary AI models, use and donate to open source ones instead. There is tons of them available. By using them, you are taking control of this technology, you are protecting your privacy and are enabling everyone to benefit equally.

We also need to wage this battle politically. Open source is a grassroots movement and when crucial legislation is being crafted we need to let our voices be heard. The government has the power to craft legislation that can kill open source. They are already doing in the US and Europe. It’s important that you take a stance whenever your state or country is making decisions about this. Sign letters and petitions that call for recognition and protection of open source principles with public access and oversight. [22]

There is tons more that I gotta cover about this. Rabbit holes that reveal the true power of billionaire lobby. For now, if you like what I do, please support me on Patreon and watch another one of my videos. I have no sponsors and my ad income doesn’t pay for my work so I am dependent on your support. Thank you.

Sources

[0]https://www.wired.com/story/runaway-ai-extinction-statement/

[1] https://www.safe.ai/work/statement-on-ai-risk

[2] https://www.cnbc.com/2024/02/09/openai-ceo-sam-altman-reportedly-seeking-trillions-of-dollars-for-ai-chip-project.html

[3] https://www.cnbc.com/2024/02/02/ai-lobbying-spikes-nearly-200percent-as-calls-for-regulation-surge.html

[4] https://www.politico.com/news/2023/10/13/open-philanthropy-funding-ai-policy-00121362

[5] https://futureoflife.org/open-letter/ai-policy-for-a-better-future-on-addressing-both-present-harms-and-emerging-threats/

[6] https://www.blumenthal.senate.gov/newsroom/press/release/blumenthal-and-hawley-announce-bipartisan-framework-on-artificial-intelligence-legislation

[7] https://www.afr.com/technology/google-brain-founder-says-big-tech-is-lying-about-ai-human-extinction-danger-20231027-p5efnz

[8] https://en.wikipedia.org/wiki/Free_and_open-source_software

[9] https://www.nytimes.com/2023/03/29/technology/ai-artificial-intelligence-musk-risks.html

[10] https://futureoflife.org/open-letter/pause-giant-ai-experiments/