Good morning, and happy Friday.

Meta is back in the hot seat — if they ever left it — following the release of a memoir called ‘Careless People,’ by former employee Sarah Wynn-Williams.

Wynn-Williams testified before a Senate Judiciary subcommittee yesterday, saying: “The greatest trick Mark Zuckerberg ever pulled was wrapping the American flag around himself and calling himself a patriot and saying he didn’t offer services in China, while he spent the last decade building an $18 billion business there.”

— Ian Krietzberg, Editor-in-Chief, The Deep View

In today’s newsletter:

• 🔬 AI for Good: A hybrid approach to new molecules

• 💰 Amazon Chief defends investments in AI; Google affirms capex

• 🏛️ OpenAI is fed up with Musk; files countersuit

AI for Good: A hybrid approach to new molecules

The major utility behind large language models (LLMs) involves their natural language format. But this same format, which enables a rather profound ease of communication between users and models, poses significant and persistent reliability challenges, specifically when it comes to scientific advancements. 

What happened: Researchers at MIT, in an attempt to leverage the raw capabilities of LLMs at a more actionable level of accuracy, designed a hybrid system designed to switch between an LLM and a graph-based machine learning algorithm. 

• Within this system, the LLM can interpret natural language queries regarding molecular design, specifications it can then send to the graph-based model for synthesis. 

• The LLM acts as a go-between, a man-to-machine translator. The result is “molecules that better matched user specifications and were more likely to have a valid synthesis plan, improving the success ratio from 5% to 35%.” 

Why it matters: LLMs struggle mightily when it comes to scientific disciplines, since the models lack any understanding of the input or output they process. Older machine learning methods tend to excel in those same scientific disciplines, but the inputs they require to do so are often highly complex. 

Here, language models are being used for … language, allowing the algorithms better suited for solving actual biological problems to shine. 

“This could hopefully be an end-to-end solution where, from start to finish, we would automate the entire process of designing and making a molecule,” co-author Michael Sun said in a statement. “If an LLM could just give you the answer in a few seconds, it would be a huge time-saver for pharmaceutical companies.” 

Big ideas. Bold strategies. AI-powered growth.

Experience it all live at ZoomInfo’s GTM 25 on May 13th at the Guggenheim Museum in New York.

This exclusive experience brings together go-to-market leaders to explore innovative GTM strategies with industry experts, including ZoomInfo Founder & CEO Henry Schuck and McKinsey & Co. Partner Todd Horst.

Connect with peers, dive deep in interactive breakout sessions, and be the first to see ZoomInfo’s groundbreaking platform innovations.

Seats are limited — reserve your spot today!

Amazon Chief defends investments in AI; Google affirms capex

Much of the debate around the viability of AI business can be summed up by this Goldman Sachs report from mid-2024, which sought to answer a simple question relevant to generative AI: “too much spend, too little benefit?” 

The investments being made by the major hyperscalers are in the hundreds of billions, due mostly to the enormous price tags associated with the special chips needed to power AI (Nvidia GPUs). But the return on that investment comes with a much longer time horizon, with Microsoft and Meta alike expecting the build-out to support revenue “over time.” 

• Then, there are the individual companies that have been experimenting the technology at steadily increasing rates. 

• A recent survey from Salesforce consultancy Coastal found that 67% of companies expect to maintain or increase their spending on AI; but only 21% reported any return on that investment. 

In the midst of this environment where adoption has proven to be both costly and challenging, and unpredictable trade wars threaten to make the infrastructure build-out even more difficult, two of the major hyperscalers re-affirmed their plans to go all-in on generative AI. 

The details: Google Chief Sundar Pichai, speaking at Google Cloud Next, said that “in 2025, we plan to invest around $75B in total capex,” a confirmation of its earlier capex expectations and a vote of confidence in AI despite mounting macroeconomic concerns. 

• Amazon, meanwhile, has committed to spending $100 billion on its infrastructure build-out. CEO Andy Jassy said in his annual shareholder letter that “generative AI is going to reinvent virtually every customer experience we know, and enable altogether new ones about which we’ve only fantasized.” 

• He said that the company’s AI revenue is growing “at triple digit (year-over-year) percentages and represents a multi-billion-dollar annual revenue run rate,” though Amazon’s specific measurement here is unclear. 

Defending the scale of the investment, Jassy added: “We continue to believe AI is a once-in-a-lifetime reinvention of everything we know, the demand is unlike anything we’ve seen before, and our customers, shareholders, and business will be well-served by our investing aggressively now.”

• ChatGPT updates: OpenAI shipped a new feature to ChatGPT on Thursday, enabling the chatbot to reference “all of your past chats to provide more personalized responses, drawing on your preferences and interests to make it even more helpful for writing, getting advice, learning and beyond.”

• Up, down, pick a direction: After its historic rally Wednesday, stocks spent Thursday in full retreat, led — once again — by the Magnificent Seven. The Dow Jones fell 2.5%, the S&P fell 3.4% and the Nasdaq fell 4.3%.

• Trump tariffs mean higher prices, big losses for Amazon sellers that source from China (CNBC).

• OpenAI gets ready to launch GPT-4.1 (The Verge).

• Data center in UK becomes proving ground for new carbon capture technology (Semafor).

• US DOGE service agreement shows $1.3 million fee — and details its mission (Wired).

• EU could tax Big Tech if Trump trade talks fail, says von der Leyen (FT).

Discover how global hiring is your competitive edge.

Hiring smarter starts with understanding salary trends.

• Get global salary data on engineers, designers, product managers and more.

• Explore top-tier talent with experience at AWS, Google, PwC, and beyond.

• Learn how to cut hiring costs and reinvest in growth.

Download the Report *

OpenAI is fed up with Musk; files countersuit 

After years of what OpenAI has termed “harassment” by the world’s richest man, Elon Musk, the startup on Wednesday fired its first return salvo: its own lawsuit.

Some quick background: Musk, himself a co-founder and investor from the earliest days of OpenAI, first sued the startup and its CEO, Sam Altman, more than a year ago. That suit alleged broadly that OpenAI wasn’t living up to its mission. Then, in June of 2024, Musk rescinded that suit, filing a new one in August that named Microsoft as an additional defendant. 

The legal battle has been intense ever since, with Musk recently amping up attempts to prevent OpenAI’s pending restructuring into a for-profit corporation. 

What happened: OpenAI said in a statement heralding the lawsuit that “Elon’s nonstop actions against us are just bad-faith tactics to slow down OpenAI and seize control of the leading AI innovations for his personal benefit.” 

“Elon’s never been about the mission,” OpenAI added. “He’s always had his own agenda.” 

• Saying that OpenAI has thus far remained “resilient” to Musk’s many attempts “to harm OpenAI,” the countersuit claims that “Musk’s actions have taken a toll,” adding that if they’re allowed to persist, OpenAI might not be able to service its mission. 

• The countersuit goes on to rehash the — at this stage — well-known history of OpenAI’s first five years, in which Musk pursued attempts to gain control of the startup, including advocating for its consumption by one of Musk’s other businesses, Tesla. 

OpenAI claims that Musk “could not abide” the startup’s success, something that inspired Musk to start his own OpenAI competitor, xAI, in 2023. 

OpenAI is accusing Musk of unfair competition and tortious interference, and further addressed and denied the claims made in Musk’s initial lawsuit. 

“The risk of future, irreparable harm from Musk’s unlawful conduct is acute, and the risk that that conduct continues is high,” the suit reads. “With every month that has passed, Musk has intensified and expanded the fronts of his campaign against OpenAI, and has proven himself willing to take ever more dramatic steps to seek a competitive advantage for xAI and to harm Altman, whom, in the words of the President of the United States, Musk ‘hates.’” 

You can read the whole filing here. 

The landscape: As OpenAI notes in the lawsuit, the company’s “capital needs have become more pressing than ever.” This desperation for a steady, powerful stream of funding has lately been on public display; last year, OpenAI secured $6.6 billion in funding alongside $4 billion in debt. Just a few months later, the company secured $40 billion in funding. 

• Both of these rounds are at least partially contingent upon OpenAI converting from its hybrid nonprofit/capped for-profit structure to a new structure. 

• The details of that new structure remain fuzzy, though OpenAI has said it would involve the capped profit turning into a public benefit corporation (PBC), and the non-profit taking a minority equity stake in the PBC. 

The lawsuit is set to go to a jury trial next spring, but OpenAI is on a tighter deadline; it must make this conversion by the end of 2025 if it wants to see the full $40 billion secured in its latest round. 

Musk’s efforts to stop the conversion — or “restructuring,” as OpenAI calls it — have gained a swell of support; a petition was recently filed asking California’s Attorney General to investigate OpenAI’s plans.

As much as this is a genuine legal response to Musk’s attacks, the lawsuit comes at a conspicuous time and frames OpenAI in a very specific, altruistic way. 

While Musk was focused on litigation, the suit claims, OpenAI was “focused on its mission.” 

That mission, according to OpenAI, keeps getting more expensive. And the investors that are willing to provide the funding for OpenAI to carry out that noble mission have “insisted” on special terms that free them from their obligations if OpenAI doesn’t complete its restructuring. 

According to the suit, the restructuring isn’t about anything else. It’s simply “designed to serve the advancement of the mission.” And if it’s halted, “OpenAI’s competitors — entities like Musk’s xAI that do not share OpenAI’s mission — will benefit.” 

But what’s actually going on is an adjustment of control. As it stands right now, OpenAI’s “primary fiduciary duty is to humanity.” But Humanity doesn’t have deep enough pockets; the proposed public benefit corporation would have “accountability to investors and employees,” and importantly, would not be beholden to the non-profit. 

It’s hard to ignore that, in 2023, Altman was fired from OpenAI by its non-profit board, a board that chose to act in pursuit and recognition of its mission. But Altman’s absence was brief; upon his return, the board’s makeup changed completely. 

The mechanism that ousted Altman in the first place will lack the power to do so again on the other end of this restructuring. 

As part of the conversion, according to Reuters, Altman himself will receive equity in OpenAI for the first time. Following its latest funding round, OpenAI is now valued at $300 billion. 

That equity stake could be worth tens of billions. 

🤔 Your thought process:

Selected Image 1 (Left):

• “Image 1 shows an unsafe but realistic way to hold a part to drill a quick hole. Yep, I've drilled a quick hole that way, my bad. In Image 2, the blocks being drilled aren't placed in any logical way on a conveyor ….”

Selected Image 2 (Right):

• “The hand position in image one seems too close to the drill bit.”

💭 A poll before you go

Thanks for reading today’s edition of The Deep View!

We’ll see you in the next one.

P.S. Enjoyed reading? Now listen! We’ve got exclusive, in-depth interviews for you on “The Deep View: Conversations” podcast every Tuesday morning. Subscribe here!

Here’s your view on Europe’s AI plans:

20% of you think Europe will dominate in AI; 20% think it won’t.

22% think it’s a real possibility, and 22% think it’s weird how hard governments are pushing AI adoption.

Never:

• “They tried this in the aviation industry and created Airbus, but that was not a cutting edge industry that is highly regulated and advances are tracked per decade and not per week like it is the case in AI. AI is moving way too fast for the Bureaucrats to keep up. Also, the whole culture is one of fear and pessimism vs the U.S. culture of optimism and innovation.”

If you want to get in front of an audience of 450,000+ developers, business leaders and tech enthusiasts, get in touch with us here.