OpenAI reveals constraints in big tech companies’ AI strategies

It’s been just two weeks since OpenAI’s cast and crew decided to celebrate the one-year anniversary of ChatGPT’s launch last year with a huge farce: firing CEO Sam Altman.

It was so much fun that my head was shaking. But before we move on, to understand what’s going on with big tech, we need to focus on its enormity, not its innovation (or even its personality). The corporate interests of a half-dozen large companies are shaping the digital transition.

With Altman back as CEO and all of OpenAI’s 700-plus staff focused on innovating around the same product as before, the only significant change has been a change from the nonprofit’s executive committee to its employees (Altman (including Mr.) only appeared to have been removed. And in true Silicon Valley fashion, two female directors were moved and replaced by two men.

But that way of thinking introduces too much of the celebrity-eyed residue of tabloid thinking into the way we cover the digital transition.

Following the funding, and on the strength of Microsoft’s US$10 billion capital investment, OpenAI has gone from a research institute working to “ensure that artificial intelligence benefits all humanity” to something like a subcontracted skunk factory: Microsoft Understand how we transitioned to an industrial research institute that employs . Its dominant “Speaking of Jump” size client.

Microsoft already owns 49% of OpenAI Global, a restricted interest subsidiary of OpenAI. Late last week, it was reported that the original and almost largest of the big tech companies would sit on the board, albeit as an observer without voting rights.

While the inevitable shift in research focus sounded good for Microsoft, the stock market ignored it, leaving the giant’s stock at the level it was before Altman was fired. Maybe what’s good for Microsoft is good for humanity…but as a for-profit company guided by the pernicious concept of shareholder value, Microsoft is a poor proxy for the interests of all of us.

This shift highlights one of Silicon Valley’s big unknowns. Can individual big technology companies innovate themselves, and us with them, toward a much different future? From what we’ve seen so far, it doesn’t seem like it. Meta, for example, has spent more than US$20 billion to pivot its traditional standby Facebook into the Metaverse, while Google has positioned itself within the broader “alphabet” and revenue sources (big Tech’s buntings are innovating (including self-driving cars). ).

OpenAI was founded by a broad group of Silicon Valley heavyweights with a nonprofit, open source ethos as a solution to the distortion of innovation by the self-interest of individual companies. In just one year, ChatGPT’s third edition soft launch has made it the fastest growing consumer technology in history with over 100 million users.

Microsoft is integrating this technology into its Bing search engine. This forced Google to rush the release of Bard and encouraged Amazon to form its own strategic alliance with Anthropic, a spinoff of Open AI.

Let’s call this the innovator’s dilemma. These companies are already making tons of money. They leveraged the opportunities of the Internet to become a top 10 company in the world. The best AI can do for them is help them stay in place, but at the expense of spending a lot of money along the way.

Or call it encitization, aka platform decline. “This is how platforms die,” said Corey Doctorow. “First, they are nice to users. Second, they abuse users to make things better for their business customers. Finally, they abuse business customers to get all the value out of them.” We will take it back to ourselves, and they will die.”

Similar to the decline of many platforms we experience every time we log on, as AI processes consumer surveillance data to market and promote what we want and don’t want; We expect the entropy of individual AI engines to skyrocket.

Or, as this week’s COP28 should remind us, it may be an environmental constraint. Artificial intelligence processes data across multiple parameters and requires extensive computing power (which is why OpenAI needed something like Microsoft’s giant Azure server farm). That computing power requires electricity, which limits the variety, scale, and applications of generative AI platforms.

As this week’s reports suggest, these challenges combine to make ChatGPT “lazy” in trying to cut corners.

Meanwhile, Silicon Valley promoters are making a fuss about what OpenAI’s outgoing Australian director Helen Toner recently called the “illusion of China’s AI capabilities” (an article now hyping China’s threat) A backlash against the regulations (read in detail in a veiled smear against her fellow director at the time, Mr. Altman). )

Open-source chat interfaces from leading technology platforms like ChatGPT (or Google’s Bard) are a significant advancement in the continued reshaping of our lives through artificial intelligence and machine learning. Can they overcome the hindrances of corporate self-interest, inevitable platform decline, and environmental constraints, or should they hope the next big thing comes from elsewhere?

Source link

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button