The Lex Newsletter: OpenAI and the case for techno-optimism

Dear reader, 

When I arrived in San Francisco in late 2018, the techlash was already in full swing. Public trust in US tech companies had been hammered by privacy violations and “fake news”. Company mottos such as Google’s “Don’t be evil” and Facebook’s “Move fast and break things” had started to sound like punchlines. 

Six years later, as I prepare to leave the city and sign off from the Lex column, San Francisco is in the grip of a new boom. SF start-ups including OpenAI and Anthropic are leading a global race in generative artificial intelligence. Tech valuations have multiplied. Look at Microsoft, the largest US company by valuation. Since 2018 its market cap has tripled to more than $3tn. Trust in tech may have declined but demand for its services and shares is at an all-time high. 

If you are a tech enthusiast you may believe that AI is about to usher in a global transformation equal to the industrial revolution. Those of a more fretful disposition will be wondering how long until it all comes crashing down. Either way, the company to watch is OpenAI. 

This week, OpenAI decided to open up the latest version of its AI chatbot to the public. The multimodal model can process visuals and audio and has a rather unnerving, giggly voice assistant. But the demonstration also served two other purposes: it kneecapped rival Google’s latest AI announcement and drew attention away from the departure of OpenAI co-founder and chief scientist Ilya Sutskever. 

Sutskever’s exit marks the final phase of OpenAI’s transformation from not-for-profit research organisation pursuing humanlike artificial general intelligence (AGI) while safeguarding humanity to multibillion-dollar commercial enterprise. It also cements the power of chief executive Sam Altman.

Go back six months and neither were certain. Sutskever had just taken part in a bizarre board coup that briefly removed Altman from OpenAI. Investors were given little explanation beyond the fact that Altman had not been “consistently candid”. The episode ended in farce with Altman’s reinstatement, supported by backer Microsoft and employees waiting to cash out their shares in the company. Sutskever recanted his actions. 

Sutskever is a pioneer in AI who was instrumental in the development of neural networks that led to the creation of large language models that power generative AI. But what turned OpenAI into an $80bn company — one of the most valuable start-ups in the world — was not just its research but its decision to repackage its breakthroughs into a free, user-friendly platform called ChatGPT. The reason it hit an annual revenue run rate of $2bn in the space of a year was the company’s decision to move quickly to sell chatbot subscriptions and access to the underlying model. 

OpenAI’s brand recognition and ability to attract attention far exceeds that of larger rivals. OpenAI says ChatGPT has 100mn weekly users. Google has not opted to provide a similar metric. Its near two-hour presentation this week showed how the company is putting AI into all of its services and wants to create personalised “agents” that can help take care of individual tasks. But as analyst Richard Windsor noted, Google is an engineering-led organisation. It is not very good at marketing its own prowess. OpenAI is.

We still don’t know how much OpenAI is spending. Nor do we know whether its pursuit of AGI is feasible. But what OpenAI has shown is that there is a demand for AI services that people are willing to pay for. That puts some substance behind AI valuations. 

The next phase we are about to enter is a tussle between big tech companies. OpenAI is reported to be in talks with Apple to add generative AI services to iPhones and other Apple products. Microsoft — which poured $13bn into OpenAI, is entitled to profits from its for-profit subsidiary and had access to its tools — is reported to be working on its own, internal AI services. OpenAI may launch a search tool to rival Google’s. 

Companies and individuals still want to be told how to use generative AI tools and what for. Is it a creativity aide or an assistant? Will it sharpen and personalise life online or are we all about to be deluged in AI-generated sludge? And is it really a threat to humanity? Answering those questions requires technical prowess of course, but it also needs showmanship. That’s something OpenAI has shown that it can excel in.

In Lex this week

• London’s quest to revitalise its stock market received a boost this week with news that Raspberry Pi, a British maker of micro computers, plans to list on the London Stock Exchange this summer. 

• Booming debt markets are offering private equity groups breathing room and the chance to avoid asset sales at cut-price valuations. 

• After years of vast and somewhat haphazard tech bets, Masayoshi Son has taken a more conservative approach to investment. As a result, SoftBank’s cash pile has grown. What are the odds that Son can resist spending it on pricey AI bets?
  

Things I enjoyed this week

Cruise was the first company to launch a self-driving taxi service in San Francisco back in 2022. I took a rather dicey trip in one of its self-driving cars last year and vowed to steer clear of robotaxis. Since then, Cruise has removed its cars from SF. Fortune magazine has a good deep dive on all the things that went wrong at the company. 

Beauty pageants may be unfashionable but hundreds of thousands of them still take place each year. However, the future of Miss USA is in question following the resignation of its latest winner and the unsettling revelation that her exit post spelt out “I AM SILENCED”.

Verge explains why lonely teenagers are making friends with AI chatbots. The information they’re sharing sounds like a privacy nightmare.  

Enjoy your weekend and thank you so much for all of the emails and reader comments on Lex tech stories over the years.

Elaine Moore
Deputy head of Lex