When scientists first sequenced the human genome in 2003, they revealed the full set of DNA instructions that make a person. But we still didn’t know what all those 3 billion genetic letters actually do.
Now Google’s DeepMind division says it’s made a leap in trying to understand the code with AlphaGenome, an AI model that predicts what effects small changes in DNA will have on an array of molecular processes, such as whether a gene’s activity will go up or down. It’s just the sort of question biologists regularly assess in lab experiments.
“We have, for the first time, created a single model that unifies many different challenges that come with understanding the genome,” says Pushmeet Kohli, a vice president for research at DeepMind.
Five years ago, the Google AI division released AlphaFold, a technology for predicting the 3D shape of proteins. That work was honored with a Nobel Prize last year and spawned a drug-discovery spinout, Isomorphic Labs, and a boom of companies that hope AI will be able to propose new drugs.
AlphaGenome is an attempt to further smooth biologists’ work by answering basic questions about how changing DNA letters alters gene activity and, eventually, how genetic mutations affect our health.
“We have these 3 billion letters of DNA that make up a human genome, but every person is slightly different, and we don’t fully understand what those differences do,” says Caleb Lareau, a computational biologist at Memorial Sloan Kettering Cancer Center who has had early access to AlphaGenome. “This is the most powerful tool to date to model that.”
Google says AlphaGenome will be free for noncommercial users and plans to release full details of the model in the future. According to Kohli, the company is exploring ways to “enable use of this model by commercial entities” such as biotech companies.
Lareau says AlphaGenome will allow certain types of experiments now done in the lab to be carried out virtually, on a computer. For instance, studies of people who’ve donated their DNA for research often turn up thousands of genetic differences, each slightly raising or lowering the chance a person gets a disease such as Alzheimer’s.
Lareau says DeepMind’s software could be used to quickly make predictions about how each of those variants works at a molecular level, something that would otherwise require time-consuming lab experiments. “You’ll get this list of gene variants, but then I want to understand which of those are actually doing something, and where can I intervene,” he says. “This system pushes us closer to a good first guess about what any variant will be doing when we observe it in a human.”
Don’t expect AlphaGenome to predict very much about individual people, however. It offers clues to nitty-gritty molecular details of gene activity, not 23andMe-type revelations of a person’s traits or ancestry.
“We haven’t designed or validated AlphaGenome for personal genome prediction, a known challenge for AI models,” Google said in a statement.
Underlying the AI system is the so-called transformer architecture invented at Google that also powers large language models like GPT-4. This one was trained on troves of experimental data produced by public scientific projects.
Lareau says the system will not broadly change how his lab works day to day but could permit new types of research. For instance, sometimes doctors encounter patients with ultra-rare cancers, bristling with unfamiliar mutations. AlphaGenome could suggest which of those mutations are really causing the root problem, possibly pointing to a treatment.
“A hallmark of cancer is that specific mutations in DNA make the wrong genes express in the wrong context,” says Julien Gagneur, a professor of computational medicine at the Technical University of Munich. “This type of tool is instrumental in narrowing down which ones mess up proper gene expression.”
The same approach could apply to patients with rare genetic disease, many of whom never learn the source of their condition, even if their DNA has been decoded. “We can obtain their genomes, but we are clueless as to which genetic alterations cause the disease,” says Gagneur. He thinks AlphaGenome could give medical scientists a new way to diagnose such cases.
Eventually, some researchers aspire to use AI to design entire genomes from the ground up and create new life forms. Others think the models will be used to create a fully virtual laboratory for drug studies. “My dream would be to simulate a virtual cell,” Demis Hassabis, CEO of Google DeepMind, said this year.
Kohli calls AlphaGenome a “milestone” on the road to that kind of system. “AlphaGenome may not model the whole cell in its entirety … but it’s starting to sort of shed light on the broader semantics of DNA,” he says.
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology.
Introducing: the Power issue
Energy is power. Those who can produce it, especially lots of it, get to exert authority in all sorts of ways.
The world is increasingly powered by both tangible electricity and intangible intelligence. Plus billionaires. The latest issue of MIT Technology Review explores those intersections, in all their forms.
Here’s just a taster of what you can expect from our latest issue:
+ Are we ready to hand AI agents the keys? We’re starting to give AI agents real autonomy, and we’re not prepared for what could happen next. Read the full story.
+ In Nebraska, a publicly owned electricity distribution system is an effective lens through which to examine the grid of the near future.
+ Cases of cancer, heart disease, and respiratory illnesses are on the rise in the area surrounding Puerto Rico’s only coal-fired power station. So why has it just been given permission to stay open for at least another seven years? Read the full story.
+ How AI is shaking up urban planning and helping make cities better.
+ Tech billionaires are making a risky bet with humanity’s future. They say they want to save humanity by creating superintelligent AI—but a new book argues that they’re steering humanity in a dangerous direction.
The Bank Secrecy Act is failing everyone. It’s time to rethink financial surveillance.
—Katie Haun is the CEO and founder of Haun Ventures, a venture capital firm focused on frontier technologies.
The US is on the brink of enacting rules for digital assets, with growing bipartisan momentum to modernize its financial system. But amid all the talk about innovation and global competitiveness, one issue has been glaringly absent: financial privacy.
As we build the digital infrastructure of the 21st century, we need to talk about not just what’s possible but what’s acceptable. That means confronting the expanding surveillance powers quietly embedded in our financial system, which today can track nearly every transaction without a warrant. Read the full story.
The must-reads
I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology.
1 Copyrighted books are fair use for AI training
According to a federal court in the US. (WP $)
+ The court compared the way AI learns to how humans consume books. (WSJ $)
+ But pirating is still illegal, apparently. (404 Media)
2 Recruiters are drowning in AI-generated résumés
Fake identities, agent-led applications, and identical résumés abound. (NYT $)
3 Extreme heat in the US is a growing threat
Alaska recently issued its first-ever heat advisory. (Vox)
+ And the heatwave is only going to intensify. (The Guardian)
+ Here’s how much heat your body can take. (MIT Technology Review)
4 Big Balls no longer works for DOGE
One of the department’s most prominent hires has resigned. (Wired $)
+ What will he do next? (NYT $)
+ DOGE’s tech takeover threatens the safety and stability of our critical data. (MIT Technology Review)
5 One of America’s best hackers is a bot
It’s the first time an AI has topped a hacking leaderboard by reputation. (Bloomberg $)
+ Cyberattacks by AI agents are coming. (MIT Technology Review)
6 Way fewer people are dying of heart attacks in the US
But deaths from chronic heart conditions are on the up. (New Scientist $)
7 TikTok’s moderators have had enough
Groups are unionizing across the world to push for better treatment. (Rest of World)
+ How an undercover content moderator polices the metaverse. (MIT Technology Review)
8 Donald Trump’s social media use is even more erratic than usual
He keeps signing off “thank you for your attention to this matter!” (The Atlantic $)
+ He’s also misspelling his name as ‘Donakd.’ (Fast Company $)
9 Finally, a use for your old smartphone
It could have a second life as a teeny tiny data center. (IEEE Spectrum)
10 AI models don’t understand Gen Alpha slang
Let him cook! (404 Media)
+ That’s not stopping youngsters from using models as advisors, though. (Fast Company $)
Quote of the day
“Humans are wired to bond, and when we feel seen and soothed—even by a machine—we connect.”
—Psychiatrist Nina Vasan explains why humans may end up falling in love with AI systems to the Wall Street Journal.
One more thing
How Wi-Fi sensing became usable tech
Wi-Fi sensing is a tantalizing concept: that the same routers bringing you the internet could also detect your movements. But, as a way to monitor health, it’s mostly been eclipsed by other technologies, like ultra-wideband radar.
Despite that, Wi-Fi sensing hasn’t gone away. Instead, it has quietly become available in millions of homes, supported by leading internet service providers, smart-home companies, and chip manufacturers.
Soon it could be invisibly monitoring our day-to-day movements for all sorts of surprising—and sometimes alarming—purposes. Read the full story.
—Meg Duff
We can still have nice things
A place for comfort, fun and distraction to brighten up your day. (Got any ideas? Drop me a line or skeet ’em at me.)
+ How to keep your cool in a heatwave.
+ Roblox fans can’t get enough of, err, gardening.
+ Kate Moss, you are the reigning queen of festival fashion.
+ A couple of intrepid brown bears managed to escape from a wildlife park in the UK—to consume a week’s worth of honey 

The US is on the brink of enacting rules for digital assets, with growing bipartisan momentum to modernize our financial system. But amid all the talk about innovation and global competitiveness, one issue has been glaringly absent: financial privacy. As we build the digital infrastructure of the 21st century, we need to talk about not just what’s possible but what’s acceptable. That means confronting the expanding surveillance powers quietly embedded in our financial system, which today can track nearly every transaction without a warrant.
Many Americans may associate financial surveillance with authoritarian regimes. Yet because of a Nixon-era law called the Bank Secrecy Act (BSA) and the digitization of finance over the past half-century, financial privacy is under increasingly serious threat here at home. Most Americans don’t realize they live under an expansive surveillance regime that likely violates their constitutional rights. Every purchase, deposit, and transaction, from the smallest Venmo payment for a coffee to a large hospital bill, creates a data point in a system that watches you—even if you’ve done nothing wrong.
As a former federal prosecutor, I care deeply about giving law enforcement the tools it needs to keep us safe. But the status quo doesn’t make us safer. It creates a false sense of security while quietly and permanently eroding the constitutional rights of millions of Americans.
When Congress enacted the BSA in 1970, cash was king and organized crime was the target. The law created a scheme whereby, ever since, banks have been required to keep certain records on their customers and turn them over to law enforcement upon request. Unlike a search warrant, which must be issued by a judge or magistrate upon a showing of probable cause that a crime was committed and that specific evidence of that crime exists in the place to be searched, this power is exercised with no checks or balances. A prosecutor can “cut a subpoena”—demanding all your bank records for the past 10 years—with no judicial oversight or limitation on scope, and at no cost to the government. The burden falls entirely on the bank. In contrast, a proper search warrant must be narrowly tailored, with probable cause and judicial authorization.
In United States v. Miller (1976), the Supreme Court upheld the BSA, reasoning that citizens have no “legitimate expectation of privacy” about information shared with third parties, like banks. Thus began the third-party doctrine, enabling law enforcement to access financial records without a warrant. The BSA has been amended several times over the years (most notoriously in 2001 as a part of the Patriot Act), imposing an ever-growing list of recordkeeping obligations on an ever-growing list of financial institutions. Today, it is virtually inescapable for everyday Americans.
In the 1970s, when the BSA was enacted, banking and noncash payments were conducted predominantly through physical means: writing checks, visiting bank branches, and using passbooks. For cash transactions, the BSA required reporting of transactions over the kingly sum of $10,000, a figure that was not pegged to inflation and remains the same today. And given the nature of banking services and the technology available at the time, individuals conducted just a handful of noncash payments per month. Today, consumers make at least one payment or banking transaction a day, and just an estimated 16% of those are in cash.
Meanwhile, emerging technologies further expand the footprint of financial data. Add to this the massive pools of personal information already collected by technology platforms—location history, search activity, communications metadata—and you create a world where financial surveillance can be linked to virtually every aspect of your identity, movement, and behavior.
Nor does the BSA actually appear to be effective at achieving its aims. In fiscal year 2024, financial institutions filed about 4.7 million Suspicious Activity Reports (SARs) and over 20 million currency transaction reports. Instead of stopping major crime, the system floods law enforcement with low-value information, overwhelming agents and obscuring real threats. Mass surveillance often reduces effectiveness by drowning law enforcement in noise. But while it doesn’t stop hackers, the BSA creates a trove of permanent info on everyone.
Worse still, the incentives are misaligned and asymmetrical. To avoid liability, financial institutions are required to report anything remotely suspicious. If they fail to file a SAR, they risk serious penalties—even indictment. But they face no consequences for overreporting. The vast overcollection of data is the unsurprising result. These practices, developed under regulations, require clearer guardrails so that executive branch actors can more safely outsource surveillance duties to private institutions.
But courts have recognized that constitutional privacy must evolve alongside technology. In 2012, the Supreme Court ruled in United States v. Jones that attaching a GPS tracker to a vehicle for prolonged surveillance constituted a search restricted by the Fourth Amendment. Justice Sonia Sotomayor, in a notable concurrence, argued that the third-party doctrine was ill suited to an era when individuals “reveal a great deal of information about themselves to third parties” merely by participating in daily life.
This legal evolution continued in 2018, when the Supreme Court held in Carpenter v. United States that accessing historical cell-phone location records held by a third party required a warrant, recognizing that “seismic shifts in digital technology” necessitate stronger protections and warning that “the fact that such information is gathered by a third party does not make it any less deserving of Fourth Amendment protection.”
The logic of Carpenter applies directly to the mass of financial records being collected today. Just as tracking a person’s phone over time reveals the “whole of their physical movements,” tracking a person’s financial life exposes travel, daily patterns, medical treatments, political affiliations, and personal associations. In many ways, because of the velocity and digital nature of today’s digital payments, financial data is among the most personal and revealing data there is—and therefore deserves the highest level of constitutional protection.
Though Miller remains formally intact, the writing is on the wall: Indiscriminate financial surveillance such as what we have today is fundamentally at odds with the Fourth Amendment in the digital age.
Technological innovations over the past several decades have brought incredible convenience to economic life. Now our privacy standards must catch up. With Congress considering landmark legislation on digital assets, it’s an important moment to consider what kind of financial system we want—not just in terms of efficiency and access, but in terms of freedom. Rather than striking down the BSA in its entirety, policymakers should narrow its reach, particularly around the bulk collection and warrantless sharing of Americans’ financial data.
Financial surveillance shouldn’t be the price of participation in modern life. The systems we build now will shape what freedom looks like for the next century. It’s time to treat financial privacy like what it is: a cornerstone of democracy, and a right worth fighting for.
Katie Haun is the CEO and founder of Haun Ventures, a venture capital firm focused on frontier technologies. She is a former federal prosecutor who created the US Justice Department’s first cryptocurrency task force. She led investigations into the Mt. Gox hack and the corrupt agents on the Silk Road task force. She clerked for US Supreme Court Justice Anthony Kennedy and is an honors graduate of Stanford Law School.
Separating AI reality from hyped-up fiction isn’t always easy. That’s why we’ve created the AI Hype Index—a simple, at-a-glance summary of everything you need to know about the state of the industry.
AI agents might be the toast of the AI industry, but they’re still not that reliable. That’s why Yoshua Bengio, one of the world’s leading AI experts, is creating his own nonprofit dedicated to guarding against deceptive agents. Not only can they mislead you, but new research suggests that the weaker an AI model powering an agent is, the less likely it is to be able to negotiate you a good deal online. Elsewhere, OpenAI has inked a deal with toymaker Mattel to develop “age-appropriate” AI-infused products. What could possibly go wrong?
The last good Instagram account
It’s a truth universally acknowledged that social media is a Bad Vibe. Thankfully, there is still one Instagram account worth following that’s just as incisive, funny, and scathing today as when it was founded back in 2016: Every Outfit (@everyoutfitonsatc). Originally conceived as an homage to Sex and the City’s iconic fashion, Every Outfit has since evolved into a wider cultural critique and spawned a podcast of the same name that I love listening to while running. Sex and the City may be over, but Every Outfit is forever.
Glorious Exploits, by Ferdia Lennon
Glorious Exploits is one of those rare books that manage to pull off being both laugh-out-loud funny and deeply moving, which is no mean feat. Set in ancient Sicily, it tells the story of unemployed potters Lampo and Gelon’s grand plan to stage the Greek tragedy Medea with a cast of defeated Athenian soldiers who’ve been imprisoned in quarries on the outskirts of Syracuse. The ancient backdrop combined with the characters’ contemporary Irish dialogue (the author was born in Dublin) makes it unlike anything I’ve ever read before; it’s so ambitious it’s hard to believe it’s Lennon’s debut novel. Completely engrossing.
Life drawing
The depressing wave of AI-generated art that’s flooded the internet in recent years has inspired me to explore the exact opposite and make art the old-fashioned way. My art teacher in college always said the best way to learn the correct proportions of the human body was to draw it in person, so I’ve started attending classes near where I live in London. Pencil and paper are generally my medium of choice. Spending a few hours interpreting what’s in front of you in your own artistic style is really rewarding—and has the added bonus of being completely screen-free. I can’t recommend it enough.
