Supremacy is everywhere; OpenAI unveils o1; Adobe releases new AI video model; Salesforce unleashes their first AI agents; Roblox uses gen AI to create games faster

The AI spending spree, in charts; AI news anchors are now a thing; tech companies offer voluntary AI commitments to the White House; Sequoia sees big money in AI apps, not models;

Sep 13, 2024

∙ Paid

In her new book Supremacy, Parmy Olson goes behind the scenes of the AI industry’s two-year rollercoaster ride which began in November 2022 with the launch of ChatGPT.

She focuses on the evolution of two founders: Sam Altman of OpenAI and Demis Hassabis of Google DeepMind, and chronicles their journeys from the early startup days where both had lofty goals of solving science and providing economic abundance to their more corporate-focused, present day predicaments, where their fortunes are closely tied to the business realities of Big Tech.

Parmy Olson for Supremacy | Edinburgh - Topping & Company Booksellers of Bath, Edinburgh, Ely, and St Andrews

While I was listening to Parmy speak to her colleague Tim O’Brien at the book launch, I couldn’t help but feel there was something missing from their conversation, given the book’s title.

You see, a lot of times the media focuses on the people at the helm of large tech companies, closely following their decisions, feuds, and, more recently, late-night social media posts. And I get why—these stories generate traffic and, to some extent, are of public interest.

But sometimes by overfocusing on the ego trips, we can miss the bigger picture and second order effects.

Let’s start with language supremacy, for example. One of the consequences of concentrating AI compute infrastructure in the US and China has been the creation of a visible gap between how well language models perform in English and Chinese compared to other languages. As these two nations invest heavily in AI research and infrastructure, their dominant languages benefit from more extensive training data and computational power. A few years ago, we had a term for this phenomenon: low-resource languages. If you asked an AI researcher for examples of such low-resource languages, they’d immediately point you to countries or communities with tens or hundreds of thousands of active speakers such as Amharic, Basque, Galician, Guarani or Høgnorsk. But now, if the current trend continues, languages with tens or hundreds of millions of speakers such as Hindi, Arabic or Spanish are at risk of becoming low resource too. The cultural impact of this trend could be profound. As AI becomes increasingly integrated into daily life, speakers of less-resourced languages may find themselves at a disadvantage in accessing AI tools and services. This could lead to a form of linguistic hegemony, where English and Chinese gain even greater regional or global influence. Additionally, the AI systems trained primarily on these languages may inadvertently encode and amplify cultural biases specific to China and the United States, potentially marginalizing other worldviews and cultural perspectives.

Secondly, national supremacy. I was born and raised in Romania. In the late 1990s and early 2000s, my country built an IT industry from scratch, with dozens of successful local tech companies that powered the economy at a time where other industries were struggling. Romania was not alone: Poland, Hungary, Estonia, India, Israel or Sweden are all small countries that punched above their weight in the global tech industry. However, unlike the previous technological wave which was driven by the internet and where having access to a computer connected to the web meant you could compete for work with anyone around the world, staying ahead in AI will be heavily influenced by access to cloud computing which can be prohibitively expensive and is presently concentrated in the hands of a few tech companies that can afford to train foundational models. These startup countries will soon find themselves in a regional Compute Deserts. Their most talented engineers and researchers will resettle in regions with more developed AI infrastructure such as the United States, Canada, China, or the Middle East. Eventually, the once-thriving tech ecosystems in Eastern Europe or the Global South will struggle to compete and will become obsolete, just like their manufacturing or automotive industries did in the 1970s and 1980s.

And now, here are the week’s news:

❤️Computer loves

Our top news picks for the week - your essential reading from the world of AI

New York Times: OpenAI Unveils New ChatGPT That Can Reason Through Math and Science
WSJ: AI Will Force a Transformation of Tech Infrastructure
The Verge: Adobe previews its upcoming text-to-video generative AI tools
WSJ: Apple Unveils an AI iPhone
ZDNet: Have a global audience? This AI video platform translates your content in one click
The Verge: Will California flip the AI industry on its head?
Wired: An AI Bot Named James Has My Old Local News Job
The Verge: Anthropic’s Mike Krieger wants to build AI products that are worth the hype
WSJ: ‘Supremacy’ Review: A Competition for Tech’s Future
New York Times: A.I. Can Now Create Lifelike Videos. Can You Tell What’s Real?
WSJ: The AI Spending Spree, in Charts

⚙️Computer does

AI in the wild: how artificial intelligence is used across industry, from the internet, social media, and retail to transportation, healthcare, banking, and more

FT: Lawyers turn to AI to expand pro bono work
Reuters: As storm Bebinca approaches, Taiwan uses AI to predict typhoon paths
BBC: How much will AI help in the next pandemic?
The Verge: Gemini’s chatty voice mode is out now for free on Android
The Verge: GitHub has started testing OpenAI’s o1-preview in GitHub Copilot
Axios: Salesforce unleashes its first AI agents
TechCrunch: Shopsense AI lets music fans buy dupes inspired by red-carpet looks at the VMAs
MIT Technology Review: Roblox is launching a generative AI that builds 3D environments in a snap
The Verge: Amazon is allowing Audible narrators to clone themselves with AI
MIT Technology Review: What impact will AI have on video game development?

🧑‍🎓Computer learns

Interesting trends and developments from various AI fields, companies and people

Keep reading with a 7-day free trial

Subscribe to Computerspeak by Alexandru Voica to keep reading this post and get 7 days of free access to the full post archives.