Will GPT-4 Lead to AGI or Dystopia?

Sam Altman is taking us to the brink.

I’ll be traveling to Singapore soon so I’m not sure how active I’ll be writing wise in the next week, just a heads up.

There is so much hype about Generative A.I. now in 2023, it cannot possibly live up to the impossible expectations set forth in the false-advertising internet we have today around the capabilities or adoption of A.I.

While I think multimodal large language models (MLLMs) will improve drastically in the mid 2020s, we are nowhere near AGI. I find myself obsessed with Microsoft’s KOSMOS-1 this weekend and its implications on the video A.I. of GPT-4 about to hit the mainstream.

I come across more articles just in ChatGPT in a day than I could possibly read, but to assume that AGI is near or just around the corner by OpenAI. But is that even close to the truth? I grow increasingly skeptical of the rhetoric even as the Venture Capital bank of Silicon Valley has suddenly imploded and Billionaires are asking the Government to bail it out.

We haven’t even created a truthful or honest internet guys. How could we regulate or keep it safe if we let an AGI lose into the world? While we ponder what’s next in LLMs, can we truly say GPT-4 will bring us a better world?

AGI is a dangerous goal for Silicon Valley, who can barely regulate itself or protect their own. The push toward humanlike artificial general intelligence (AGI) however will require equipping such models with additional capabilities — and multimodal perception is an essential next step. But as robots themselves develop these capabilities, are we building the foundations for dystopia?

To embody MLLMs into the world, is to set a dangerous momentum of consequences we cannot even imagine in 2023. Perhaps we should stop to consider what we are doing with A.I., before it’s too late to reverse the exponential fever of funding that will build more and better. I cannot help but sense I’m living through an end-of-times moment, that is, the beginning of dystopia. The middle class is being crushed by inflation and debt, and A.I. is the supposed savior? How do those sci-fi novels usually end up?

I enjoy writing about A.I., but what will be the cost of the world we are creating with it increasingly at the helm? What will stop Google, OpenAI, Tesla or Microsoft from placing MLLMs inside of robots at scale? And what might that do to humanity if a general purpose robot becomes cheaper to make than it is to hire a person? Is that really the utopia we want for our children?

Progress doesn’t always bring the outcome we had hoped. Smart phones aren’t doing us any favors for our mental health or the most important thing in well-being, real social connectivity. But the machine demands our attention, and so does the cash-cow of digital advertising. Now those funds will be used to develop A.I. systems of further enhancement and entrapment, whatever the brain-computer-interface turns out to be. Likely a neuro-mesh with A.I. itself.

As techno-optimistic as I try to be, I can’t help feeling we might be letting our descendants down with how we are building the future of A.I. We are heralding the tools of automation and wealth-inequality at such a scale I don’t believe our democracies or our versions of capitalism will survive it. The rise of Generative A.I. and the fall of Silicon Valley Bank colliding in 2023, is not a coincidence that’s lost on me. And Silicon Valley is telling us to bail out its favorite bank.

Meanwhile OpenAI is shrouded in secrecy and heralding AGI as if this is good for society, humanity and free-will? If I cannot even prompt ChatGPT (and it’s dozens of clones) well, I’m actually at a significant disadvantage to even keep up. I don’t even believe in the capability of this being empowering A.I. The woke left-leaning biased foundational models are a funnel of supposedly good information.

Microsoft Germany CTO, Andreas Braun, confirmed that GPT-4 is coming within a week of March 9, 2023 and that it will be multimodal. Will multimodal Artificial Generative Intelligence (MAGI) be good for monopoly Capitalism, or will it just enable companies like Microsoft, Google, Amazon and Apple to become even more powerful crushing all that opposes them? Even as LinkedIn becomes some cringe version of the TikTok of business professional posts. Do I now want GPT-4 generated videos on LinkedIn to improve my productivity?

LLMs with multimodal perception will be better equipped to acquire commonsense knowledge beyond the information they glean from text alone; and that this perception enrichment will facilitate LLM applications in new domains such as robotics and document intelligence. Do I want to live in a world full of robots, chatbots and potentially even an AGI to thwart, compete with and serve me? Smarter than my cat and only bound to get even more capable in the ever more sophisticated machine driven economy?

GPT-4 might be the biggest step towards AGI we’ve had in many decades, but that doesn’t mean that it will lead to a better world. In fact, the Ukraine war and the response to the pandemic has shown humanity isn’t ready for more dangerous tools less a lot of people will get seriously hurt.