This is the latest issue of The Download, our weekday newsletter delivering a daily update on the happenings in technology.
The Situation with AI: Entering the economic singularity
—David Rotman and Richard Waters
Any significant new technology tends to have inconsistent levels of adoption, but generative AI has been particularly disparate. This variability complicates the ability to predict its potential effects on individual enterprises, much less on overall productivity within the economy.
On one hand, AI coding assistants have transformed software development. Conversely, many companies are witnessing minimal to no advantages from their early investments.
This discrepancy has fueled skepticism among those who argue that generative AI, being a probabilistic technology susceptible to generating errors, will lack a profound influence on business. For those knowledgeable in tech history, however, the absence of immediate results is typical. Read the complete article.
For MIT Technology Review subscribers, join David and Richard with our editor in chief, Mat Honan, for an exclusive discussion diving into current market dynamics live on Tuesday, December 9 at 1pm ET. Register here!
The State of AI is our exclusive collaboration for subscribers between the Financial Times and MIT Technology Review exploring how AI is altering global influence.Sign up to receive future editions every Monday.
The must-reads
I’ve searched the internet to bring you today’s most entertaining/important/alarming/intriguing technology stories.
1 DeepSeek has launched two new experimental AI models
DeepSeek-V3.2 aims to compete with OpenAI’s GPT-5 in reasoning abilities. (Bloomberg $)
+ Here’s how DeepSeek reduces its models’ computational load. (VentureBeat)
+ The company has achieved these outcomes despite limited access to advanced processors. (SCMP $)
2 OpenAI has declared a “code red” to its employees
This is a rallying cry to enhance ChatGPT, or risk falling behind. (The Information $)
+ Google and Anthropic are closely pursuing OpenAI. (FT $)
+ Advertising and other projects will be delayed to focus on this new direction. (WSJ $)
3 Recognizing the signs of an AI bubble burst
These are indicators to watch for. (Economist $)
+ The economy could face severe consequences if and when it occurs. (Axios)
+ We remain uncertain about how the AI investment boom is funded. (The Guardian)
4 Certain US states are prohibiting AI from discriminating against individuals
California is the latest to empower workers to combat algorithmic biases. (WP $)
5 This AI startup is pursuing a post-transformer future
The transformer architecture is central to the current AI surge—but Pathway is innovating something new. (WSJ $)
+ Envisioning the next phase of AI innovation. (IEEE Spectrum)
6 India is insisting that smartphone manufacturers include a government app
Privacy advocates deem this an unacceptable invasion of privacy. (FT $)
+ India’s tech manpower is seeking opportunities beyond the US. (Rest of World)
7 College students are eager to enroll in AI studies
AI is now the second-largest major at MIT, following computer science. (NYT $)
+ Major AI companies aspire to dominate the educational space. (MIT Technology Review)
8 America’s musical legacy is under serious threat
Much of it resides on studio tapes, which are gradually degrading. (NYT $)
+ The urgency to preserve our digital existence from an impending dark age. (MIT Technology Review)
9 Celebrities are increasingly opposing AI
However, fans continue to feature them in questionable videos regardless. (The Verge)
10 Samsung has launched its inaugural tri-folding phone
But will there be demand for it? (Bloomberg $)
+ It will retail for over $2,000 when it becomes available in South Korea. (Reuters)
Quote of the day
“The Chinese will not slow down. They will dominate.”
—Michael Lohscheller, CEO of the Swedish electric vehicle manufacturer Polestar, tells the Guardian why Europe must continue its strategy to prohibit the production of new petrol and diesel vehicles by 2035.
One more thing
Inside Amsterdam’s high-stakes initiative to develop equitable welfare AI
Amsterdam believed it was on a positive path. City officials from the welfare department thought they could create technology to prevent fraud while safeguarding citizens’ rights. Following emerging best practices and investing considerable time and finances in a project that ultimately processed live welfare applications, they discovered during their pilot that the system they designed still fell short of fairness and effectiveness. What went wrong?
Lighthouse Reports, MIT Technology Review, and the Dutch publication Trouw have gained unprecedented insight into the system to investigate. Read about our findings.
—Eileen Guo, Gabriel Geiger & Justin-Casimir Braun
We can still enjoy nice things
A space for comfort, entertainment, and diversion to enhance your day. (Have suggestions? Send me a message or hit me up.)
+ Consider this: a great holiday movie doesn’t need to revolve around Christmas at all.
+ Perhaps we should reevaluate the importance of book covers.
+ Happy birthday to Ms Britney Spears, still reigning as the princess of pop at 44!
+ The compelling psychology behind our love for traveling so much.
