The wish world of Silicon Valley

This week: Why does Silicon Valley seem to love Universal Basic Income? AI getting better at troubleshooting, and hacking. Anthropic is off the hook, and computing power created the new divide.

Last week's Kludder was mostly about how AI will take our jobs. This week we look at the solutions the AI leaders themselves have proposed if we all end up unemployed. Universal basic income seems to be the favorite.

Universal basic income is a guaranteed income for all members of a community, such as a country. In step with the development of artificial intelligence, the world's richest tech entrepreneurs have tested the idea of giving people money without conditions. And the idea of universal basic income has powerful supporters, including Elon Musk, OpenAI founder Sam Altman, Twitter founder Jack Dorsey, and Salesforce CEO Marc Benioff.

OpenResearch is funded by OpenAI founder Sam Altman with 60 million dollars. Last year, the research institution published results from a comprehensive universal basic income experiment called The Unconditional Income Study.

For me, these findings raise more questions than they answer.

The final proof

3,000 people in Texas and Illinois with annual incomes under $30,000 participated in the study over a three-year period. One-third received $1,000 per month, while the control group received $50 monthly.

Cash transfers probably do less to improve people's lives than proponents thought they would. On the other hand, they probably don't have the harmful effects that critics were worried about.

This is how Sarah Miller, economist at the University of Michigan and co-author of the final reports, concludes to the New York Times.

The sum did not become the catalyst for increased quality of life that was hoped for. Recipients spent an average of $310 more per month, with food ($67), rent ($52), and transportation ($50) as the largest expense items. At the same time, their net worth did not increase - the money was often offset by recipients increasing their debt levels through car loans and mortgages.

Sam Altman has long been a believer of universal basic income, also known as guaranteed income, in some form or other. It seems to spring from a conviction Altman's got where artificial intelligence replaces most jobs. The OpenAI CEO wrote this in 2016, when he was head of startup incubator Y Combinator:

In 50 years, I think it will seem ridiculous that we used the fear of not being able to eat as a way to motivate people. - Sam Altman

Why so generous?

The dream of universal basic income has truly gained a foothold among the tech bros in California: In Santa Clara County, there are now nine pilot programs through Destination: Home. A total of 950 people receive about $1,000 monthly, and companies such as Google, Cisco, and Apple contribute one-third of the $26 million budget.

The result from the research project wasn't a knockout victory for Altman. And yet, if the results are so mediocre, why do tech billionaires continue to pump millions into the research?

Karl Widerquist, philosophy professor at Georgetown University in Qatar, believes it's about time the technology sector finally takes responsibility.

They’re using our data to create their products and haven’t paid us back. And they’re the ones saying they’ll disrupt the economy and put people out of work. - Karl Widerquist

I don't disagree. But the question is whether or not universal basic income is the way to go.

The Norwegian discourse

Here in Norway, there's also debate around universal basic income. Chief economist Jan Ludvig Andreassen at Eika Group sees possibilities for regional policy. In a press release (Norwegian), Andreassen points out that universal basic income provides certainty about economic security, which can be used to maintain settlement patterns and get people to move to Northern Norway.

The list to lure people to Northern Norway is getting comical: Cancel student loans, give them free kindergarten, and now universal basic income! I suggest focusing on getting interesting businesses that provide good career opportunities to relocate. That way people will follow. But this is a completely different topic.

Andreassen also claims that universal basic income "most often ends in laziness and substance abuse." It's questionable whether the Eika economist has read the results of Altman's and OpenResearch's study. It shows that recipients did not spend the money on substances, but rather on basic needs and bills.

The new currency

Despite his enthusiasm for universal basic income, Altman also has another currency in mind. He envisions a world driven not by money, but by computational power.

I wonder if the future looks more like universal basic compute than universal basic income. - Sam Altman to All-In Podcast

The idea is that when OpenAI builds general artificial intelligence, everyone should be able to get a small share and access to it. We can use this, sell it on, or donate it to research. And Altman may have been thinking about this for a while. On OpenAI's website it states, according to Politico:

"It would be wise to view any investment in OpenAI Global, LLC in the spirit of a donation... with the understanding that it may be difficult to know what role money will play in a world with artificial general intelligence."

Maybe we're already at a place where data is the real currency. Senior economist at the Foundation for American Innovation, Samuel Hammond, believes data seems to be a bigger key to success than money for startup companies. Companies like Nvidia can, for example, entice with computational power and microchips, something that can be more valuable than fresh capital in the cash register:

Companies are literally raising capital in the form of compute, as if it were currency. - Samuel Hammond.

Is this the future?

Universal basic income sounds good on paper. I could spend my days going swimming, writing Kludder, and bouldering. A kind of dream life for me. But what happens to us when we no longer have a job to go to, or no longer meet people we otherwise wouldn't have encountered?

I fear universal basic income could create a new class divide: Those who create, and those who consume. And OpenResearch's study shows that universal basic income is not the miracle cure supporters hoped for. Yet. Nor is it the catastrophe critics feared.

But falling into a debate exclusively about whether universal basic income is the wrong way to go. If Norway, and Europe, are to have any say in the years to come, the focus needs to be on what kind of society we want to build in the future. About how technology can pull us in the right direction, without us losing ourselves in a world of AI-generated news and AI-driven businesses. Because artificial intelligence is going to change fundamental conditions for work and income.

I don't know about you, but I don't want to leave the work force just yet.

Ask me again in 35 years.

Can reinforce prejudices

Ever since language models became part of our lives, I've been puzzled that they always seem to agree with me. Unless I specifically ask for criticism and order it to tone down the sycophancy, my ideas will always be perceived as the world's best. A group of researchers has looked at exactly this.

This type of behavior is seen as problematic. Not only can the learning outcome be reduced, but it can reinforce biases and prejudices. It can also hinder critical thinking and push you toward the wrong type of decisions. An interview in the Wall Street Journal with Malihe Alikhani, professor in AI, provides a deep dive into the problems that arise when ChatGPT just wants to flatter.

The AI divide

In Texas, Sam Altman is building a data center the size of Central Park in New York. $60 billion is being poured into the project.

This New York Times article shows how artificial intelligence is creating a new divide: Those with access to massive amounts of computational power who can build cutting-edge AI systems, and those who don't have it.

Artificial intelligence has created a new digital divide, fracturing the world between nations with the computing power for building cutting-edge A.I. systems and those without. The split is influencing geopolitics and global economics, creating new dependencies and prompting a desperate rush to not be excluded from a technology race that could reorder economies, drive scientific discovery and change the way that people live and work. - Excerpt from the article

Protector and hacker

Researchers at UC Berkeley have tested how good AI models have become at detecting errors and flaws in code. Through using a new measurement method called CyberGum, it emerged that the models themselves discovered critical errors that had not been detected until now.

But at the same time as AI becomes better at debugging, they also become better at hacking. Thus we're heading into a time where AI becomes crucial for cybersecurity, while at the same time being the main tool for committing cybercrime. Wired goes in-depth on the topic, but if you don't have a subscription, I can recommend this blog post written by several security experts.

Anthropic got away

A few weeks ago I wrote about the online newspapers struggling in the new search economy. This week came the verdict in a lawsuit between Anthropic, the company behind the language model Claude, and the authors.

A federal judge in San Francisco has acquitted the AI company. The reasoning is that Anthropic used the books by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson in a fair way to train Claude. The judge compared Anthropic's use of books to a reader who wants to become an author and who uses works "not to run past and replicate or replace them" but to "create something different".