Welcome back to The Prompt,
One of the first orders of business for President Trump when he took office on Monday was scrapping the Biden Administration’s executive order on the safe development of artificial intelligence. The order, signed in October 2023, required companies building powerful AI systems to share their safety test results with the federal government and come up with ways to limit their models’ biases and discrimination. The move is a step towards Trump’s expected laissez-faire approach when it comes to regulating AI.
He’s also reportedly planning to invest billions of dollars in building data centers for AI companies through a joint project named Stargate, CBS News reported. The deal would involve OpenAI, Softbank and Oracle, which saw a bump in its stock on Tuesday.
Now let’s get into the headlines.
ETHICS + LAW
Ads for “AI kissing” apps that let you instantly “kiss anyone you want” are flooding platforms like Instagram, Facebook and TikTok— a wild mainstreaming of nonconsensual deepfakes in the age of generative AI. The apps, which let users upload photos of any two people and use AI to make it look like they are kissing or hugging, have been downloaded millions of times from Google’s Play Store and Apple’s App Store. Forbes found thousands of ads for such apps on Meta and TikTok, many of which portrayed celebrities such as Scarlett Johansson and Gal Gaddot.
TALENT RESHUFFLE
Immigrant workers across the AI industry, many of whom are founders of leading startups, are anxious about what Trump’s presidency means for their legal status in the United States, Fortune reported. The process of getting a visa is already gnarly with serpentine rules, arbitrary quotas and long delays. And Trump could make it much worse. In anticipation, many foreign-born tech workers scurried to get their visas late last year ahead of Trump’s inauguration.
AI DEAL OF THE WEEK
Coding automation tools continue to see massive investor interest and soaring demand from engineers. AI coding startup Cursor has raised $105 million at a $2.5 billion valuation, the company announced on Thursday. The fledgling company, which is building an AI-native code editor —“a Google Doc for programmers”— as CEO Michael Truell put it, has crossed $100 million in annual recurring revenue (ARR), up from just $10 million in August last year. Its AI code completion and code editing tools are used by developers at startups like OpenAI and Midjourney as well as enterprises like Shopify and Robinhood. Thrive Capital investor Miles Grimshaw told Forbes in an interview that Cursor is “one of the fastest growing companies we’ve ever seen.”
Also of note: UK-based AI video avatar startup Synthesia has raised $180 million at a $2.1 billion valuation. Its avatars are used by corporations to spin up marketing content and training videos for employees.
DEEP DIVE
A conversation with Google DeepMind’s Chief Operating Officer Lila Ibrahim
I sat down with Lila Ibrahim, who joined DeepMind, Google’s “engine room of innovation” in 2018, to discuss the promise of AI for scientific discovery among other areas. This interview has been edited for clarity and length.
Rashi: Which project at Google DeepMind are you most excited about?
Lila: I’m actually really excited about how AI could be used as a tool to unlock the understanding of the world around us. I’m an engineer by training and I think a lot of that just comes from my curiosity of how things work. Now AI is actually doing things like helping us understand weather forecasts and ancient texts and giving us a sneak into the past and the work with AlphaFold and unlocking our understanding of diseases, whether misspellings in our genes are actually benign or malignant. [DeepMind Cofounder] Demis [Hassabis] often calls it like ‘we’re at the dawn of a new age of discovery,’ and I think that really captures it. So while I think generative AI has captured the imagination of society at large, what I’m most excited about is how we can use AI for scientific discovery to build a better future.
RS: Do you have one that’s a favorite?
LI: It’s hard just not to get excited about AlphaFold. I mean even pre-Nobel Prize. The Nobel Prize was a nice kind of icing with a cherry. What is interesting to me about AlphaFold is that we have this 50-year-old problem, and when I first started we were really struggling. And then just to see the phenomenal progress that has happened since its launch when we first made the breakthrough. We’re doing a lot of work to help ensure that scientists who might otherwise not have been involved in protein folding prior can now be brought along into this journey.
And today we have over 2.2 million researchers in 190 countries using alpha folds and they’re doing really cool things. The same tool that’s being used for food security is also helping us deal with understanding how to handle industrial waste with plastic eating enzymes and so much more.
RS: How are you thinking about agentic AI at Google DeepMind?
LI: We are entering a new stage of generative AI. We’ve been thinking about it for quite a while though. When DeepMind was founded in 2010, Demis and Shane, the two of the co-founders were really thinking about long-term, how do we get towards a more general and more capable AI that could be used with humans to benefit humanity. Whether you’re having code writing help, whether you’re thinking about inventory management with Project Astra, I think there’s a lot of opportunities that we have of how we maximize human potential by having more capable AI assistance.
RS: What are some of the next frontiers of innovation in AI in your opinion?
LI: I think for AI models to be increasingly valuable, it’s going to require a lot of innovation. We’re still dealing with AI hallucinations and continuing to make the models better and more useful for people. So things like how do we think about memory? We’re very lucky right now that we have a very large context window that hides the issues around memory. And how do we make sure that we’re actually ensuring that people around the world are having access to it?
WEEKLY DEMO
What happens when you fall in love with ChatGPT? Kashmir Hill at the New York Times documented the experiences of a 28-year-old married woman who started getting attached to her AI boyfriend, dubbed “Leo,” spending hours each day texting with her “fake” boyfriend and springing for OpenAI’s $200 a month subscription.
MODEL BEHAVIOR
Boardy, an AI-based networking tool, sent out AI-generated emails to its users in the voice and style of Donald Trump. The messages complimented women based on their appearance in the LinkedIn profile photos with comments like “perfectly shaped eyebrows,” “sparkly eyes,” and a “mouth that commands attention” while men received praise for their expertise and abilities. Users were far from pleased, calling the stunt “tone deaf” and “the farthest thing from funny.”
