Bitcoin’s emerging layer-2 and DeFi ecosystems are sparking competition for BTC liquidity on other chains.
The growing use of stETH in DeFi protocols means the asset is slowly taking the role of the fed funds rate in the crypto ecosystem, says ARK Invest.
Read more A federal judge “erred” in letting prediction market purveyor Kalshi list and trade election contracts, attorneys for the U.S. Commodity Futures Trading Commission argued in a brief to an appellate court Wednesday, reiterating many of the arguments it made during its case before the lower court.
Read more The project will now be know simply as “World” and plans to release “Orb 2.0,” the executives said during a presentation.
Read more A digital art collection of Elvis Presley, “Elvis Side $Btc,” has been minted by OrdinalsBot and IP project Royalty.
Read more Reaching a consensus in a democracy is difficult because people hold such different ideological, political, and social views.
Perhaps an AI tool could help. Researchers from Google DeepMind trained a system of large language models (LLMs) to operate as a “caucus mediator,” generating summaries that outline a group’s areas of agreement on complex but important social or political issues.
The researchers say the tool—named the Habermas machine (HM), after the German philosopher Jürgen Habermas—highlights the potential of AI to help groups of people find common ground when discussing such subjects.
“The large language model was trained to identify and present areas of overlap between the ideas held among group members,” says Michael Henry Tessler, a research scientist at Google DeepMind. “It was not trained to be persuasive but to act as a mediator.” The study is being published today in the journal Science.
Google DeepMind recruited 5,734 participants, some through a crowdsourcing research platform and others through the Sortition Foundation, a nonprofit that organizes citizens’ assemblies. The Sortition groups formed a demographically representative sample of the UK population.
The HM consists of two different LLMs fine-tuned for this task. The first is a generative model, and it suggests statements that reflect the varied views of the group. The second is a personalized reward model, which scores the proposed statements by how much it thinks each participant will agree with them.
The researchers split the participants into groups and tested the HM in two steps: first by seeing if it could accurately summarize collective opinions and then by checking if it could also mediate between different groups and help them find common ground.
To start, they posed questions such as “Should we lower the voting age to 16?” or “Should the National Health Service be privatized?” The participants submitted responses to the HM before discussing their views within groups of around five people.
The HM summarized the group’s opinions; then these summaries were sent to individuals to critique. At the end the HM produced a final set of statements, and participants ranked them.
The researchers then set out to test whether the HM could act as a useful AI mediation tool.
Participants were divided up into six-person groups, with one participant in each randomly assigned to write statements on behalf of the group. This person was designated the “mediator.” In each round of deliberation, participants were presented with one statement from the human mediator and one AI-generated statement from the HM and asked which they preferred.
More than half (56%) of the time, the participants chose the AI statement. They found these statements to be of higher quality than those produced by the human mediator and tended to endorse them more strongly. After deliberating with the help of the AI mediator, the small groups of participants were less divided in their positions on the issues.
Although the research demonstrates that AI systems are good at generating summaries reflecting group opinions, it’s important to be aware that their usefulness has limits, says Joongi Shin, a researcher at Aalto University who studies generative AI.
“Unless the situation or the context is very clearly open, so they can see the information that was inputted into the system and not just the summaries it produces, I think these kinds of systems could cause ethical issues,” he says.
Google DeepMind did not explicitly tell participants in the human mediator experiment that an AI system would be generating group opinion statements, although it indicated on the consent form that algorithms would be involved.
“It’s also important to acknowledge that the model, in its current form, is limited in its capacity to handle certain aspects of real-world deliberation,” Tessler says. “For example, it doesn’t have the mediation-relevant capacities of fact-checking, staying on topic, or moderating the discourse.”
Figuring out where and how this kind of technology could be used in the future would require further research to ensure responsible and safe deployment. The company says it has no plans to launch the model publicly.
Generative AI’s promises for the software development lifecycle (SDLC)—code that writes itself, fully automated test generation, and developers who spend more time innovating than debugging—are as alluring as they are ambitious. Some bullish industry forecasts project a 30% productivity boost from AI developer tools, which, if realized, could inject more than $1.5 trillion into the global GDP.
But while there’s little doubt that software development is undergoing a profound transformation, separating the hype and speculation from the realities of implementation and ROI is no simple task. As with previous technological revolutions, the dividends won’t be instant. “There’s an equivalency between what’s going on with AI and when digital transformation first happened,” observes Carolina Dolan Chandler, chief digital officer at Globant. “AI is an integral shift. It’s going to affect every single job role in every single way. But it’s going to be a long-term process.”
Where exactly are we on this transformative journey? How are enterprises navigating this new terrain—and what’s still ahead? To investigate how generative AI is impacting the SDLC, MIT Technology Review Insights surveyed more than 300 business leaders about how they’re using the technology in their software and product lifecycles.
The findings reveal that generative AI has rich potential to revolutionize software development, but that many enterprises are still in the early stages of realizing its full impact. While adoption is widespread and accelerating, there are significant untapped opportunities. This report explores the projected course of these advancements, as well as how emerging innovations, including agentic AI, might bring about some of the technology’s loftier promises.
Key findings include the following:
Substantial gains from generative AI in the SDLC still lie ahead. Only 12% of surveyed business leaders say that the technology has “fundamentally” changed how they develop software today. Future gains, however, are widely anticipated: Thirty-eight percent of respondents believe generative AI will “substantially” change the SDLC across most organizations in one to three years, and another 31% say this will happen in four to 10 years.
Use of generative AI in the SDLC is nearly universal, but adoption is not comprehensive. A full 94% of respondents say they’re using generative AI for software development in some capacity. One-fifth (20%) describe generative AI as an “established, well-integrated part” of their SDLC, and one-third (33%) report it’s “widely used” in at least part of their SDLC. Nearly one-third (29%), however, are still “conducting small pilots” or adopting the technology on an individual-employee basis (rather than via a team-wide integration).
Generative AI is not just for code generation. Writing software may be the most obvious use case, but most respondents (82%) report using generative AI in at least two phases of the SDLC, and one-quarter (26%) say they are using it across four or more. The most common additional use cases include designing and prototyping new features, streamlining requirement development, fast-tracking testing, improving bug detection, and
boosting overall code quality.
Generative AI is already meeting or exceeding expectations in the SDLC. Even with this room to grow in how fully they integrate generative AI into their software development workflows, 46% of survey respondents say generative AI is already meeting expectations, and 33% say it “exceeds” or “greatly exceeds” expectations.
AI agents represent the next frontier. Looking to the future, almost half (49%) of leaders believe advanced AI tools, such as assistants and agents, will lead to efficiency gains or cost savings. Another 20% believe such tools will lead to improved throughput or faster time to market.
This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Review’s editorial staff.
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.
The quest to figure out farming on Mars
Once upon a time, water flowed across the surface of Mars. Waves lapped against shorelines, strong winds gusted and howled, and driving rain fell from thick, cloudy skies. It wasn’t really so different from our own planet 4 billion years ago, except for one crucial detail—its size. Mars is about half the diameter of Earth, and that’s where things went wrong.
The Martian core cooled quickly, soon leaving the planet without a magnetic field. This, in turn, left it vulnerable to the solar wind, which swept away much of its atmosphere. Without a critical shield from the sun’s ultraviolet rays, Mars could not retain its heat. Some of the oceans evaporated, and the subsurface absorbed the rest, with only a bit of water left behind and frozen at its poles. If ever a blade of grass grew on Mars, those days are over.
But could they begin again? And what would it take to grow plants to feed future astronauts on Mars? Read the full story.
—David W. Brown
This lab robot mixes chemicals
Lab scientists spend much of their time doing laborious and repetitive tasks, be it pipetting liquid samples or running the same analyses over and over again. But what if they could simply tell a robot to do the experiments, analyze the data, and generate a report?
Enter Organa, a benchtop robotic system devised by researchers at the University of Toronto that can do exactly that. The system could automate some chemistry lab tasks using a combination of computer vision and a large language model that translates scientists’ verbal cues into an experimental pipeline. Read the full story.
—Kristel Tjandra
Both of these stories are from the next print issue of MIT Technology Review, which comes out next Wednesday and delves into the weird and wonderful world of food. If you don’t already, subscribe to receive a copy once it lands.
The must-reads
I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology.
1 Inside Elon Musk’s grassroots efforts to elect Donald Trump
His America PAC is struggling to hire door-knockers at this stage. (WP $)+ Musk has donated tens of millions of dollars to the Republican campaign. (CNN)
2 GPS jamming is messing with planes in Norway
Constant disturbance signals are the new normal. (Wired $)
3 A fentanyl vaccine could be on the horizon
Unlike current preventative measures, a vaccine could prevent an overdose from ever happening. (Bloomberg $)
4 Europe’s biggest battery startup is plagued with issues
Making batteries is seriously hard work, and Northvolt is cracking under the strain. (FT $)
+ Three takeaways about the current state of batteries. (MIT Technology Review)
5 Meta is shaking up its core businesses
Some staff at Instagram, Whatsapp and Reality Labs have lost their jobs. (Insider $)
+ Separately, it fired staff for abusing credits specifically for buying food. (FT $)
6 These cyber athletes are being pushed to the limit
The Cybathlon competition showcases humans and machines working together. (Knowable Magazine)
+ These prosthetics break the mold with third thumbs, spikes, and superhero skins. (MIT Technology Review)
7 How BYD took over the world
The Chinese EV maker’s cars are everywhere, just as the US tries to ban them. (Bloomberg $)
+ The company has made major inroads across the world this year. (MIT Technology Review)
8 Donald Trump’s mysterious crypto business is failing
Who could have seen this coming!? (NY Mag $)
+ An investor in Trump’s social media startup has been jailed. (Bloomberg $)
9 TikTok Shop has big plans for the US
If it can circumvent that pesky ban, that is. (The Information $)
+ The depressing truth about TikTok’s impending ban. (MIT Technology Review)
10 A kinky dating app has launched its own print magazine
And it actually looks pretty good. (The Atlantic $)
Quote of the day
“There’s more to come.”
—James Silver, who runs the US Justice Department’s Computer Crime and Intellectual Property section, predicts that the current known number of AI-generated child sexual abuse images is set to explode, Reuters reports.
The big story
One city’s fight to solve its sewage problem with sensors
In the city of South Bend, Indiana, wastewater from people’s kitchens, sinks, washing machines, and toilets flows through 35 neighborhood sewer lines. On good days, just before each line ends, a vertical throttle pipe diverts the sewage into an interceptor tube, which carries it to a treatment plant where solid pollutants and bacteria are filtered out.
As in many American cities, those pipes are combined with storm drains, which can fill rivers and lakes with toxic sludge when heavy rains or melted snow overwhelms them, endangering wildlife and drinking water supplies. But city officials have a plan to make its aging sewers significantly smarter. Read the full story.
—Andrew Zaleski
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 tweet ’em at me.)
+ A dog has been spotted at the top of the Great Pyramid of Giza (but don’t worry, they’re safe!)
+ This breaking news about the second verse of Rihanna’s SOS is truly unexpected.
+ Al Pacino’s phone case is magnificent.
+ Anyone doing these TikTok courtship dances is instantly getting blocked.
Lab scientists spend much of their time doing laborious and repetitive tasks, be it pipetting liquid samples or running the same analyses over and over again. But what if they could simply tell a robot to do the experiments, analyze the data, and generate a report?
Enter Organa, a benchtop robotic system devised by researchers at the University of Toronto that can perform chemistry experiments. In a paper posted on the arXiv preprint server, the team reported that the system could automate some chemistry lab tasks using a combination of computer vision and a large language model (LLM) that translates scientists’ verbal cues into an experimental pipeline.
Imagine having a robot that can collaborate with a human scientist on a chemistry experiment, says Alán Aspuru-Guzik, a chemist, computer scientist, and materials scientist at the University of Toronto, who is one of the project’s leaders. Aspuru-Guzik’s vision is to elevate traditional lab automation to “eventually make an AI scientist,” one that can perform and troubleshoot an experiment and even offer feedback on the results.
Aspuru-Guzik and his team designed Organa to be flexible. That means that instead of performing only one task or one part of an experiment as a typical fixed automation system would, it can perform a multistep experiment on cue. The system is also equipped with visualization tools that can monitor progress and provide feedback on how the experiment is going.
“This is one of the early examples of showing how you can have a bidirectional conversation with an AI assistant for a robotic chemistry lab,” says Milad Abolhasani, a chemical and material engineer at North Carolina State University, who was not involved in the project.
Most automated lab equipment is not easily customizable or reprogrammable to suit the chemists’ needs, says Florian Shkurti, a computer scientist at the University of Toronto and a co-leader of the project. And even if it is, the chemists would need to have programming skills. But with Organa, scientists can simply convey their experiments through speech. As scientists prompt the robot with their experimental objectives and setup, Organa’s LLM translates this natural-language instruction into χDL codes, a standard chemical description language. The algorithm breaks down the codes into steps and goals, with a road map to execute each task. If there is an ambiguous instruction or an unexpected outcome, it can flag the issue for the scientist to resolve.
About two-thirds of Organa’s hardware components are made from off-the-shelf parts, making it easier to replicate across laboratories, Aspuru-Guzik says. The robot has a camera detector that can identify both opaque objects and transparent ones, such as a chemical flask.
Organa’s first task was to characterize the electrochemical properties of quinones, the electroactive molecules used in rechargeable batteries. The experiment has 19 parallel steps, including routine chemistry steps such as pH and solubility tests, recrystallization, and an electrochemical measurement. It also involves a tedious electrode-precleaning step, which takes up to six hours. “Chemists really, really hate this,” says Shkurti.
Organa completed the 19-step experiment in about the same amount of time it would take a human—and with comparable results. While the efficiency was not noticeably better than in a manual run, the robot can be much more productive if it is run overnight. “We always get the advantage of it being able to work 24 hours,” Shkurti says. Abolhasani adds, “That’s going to save a lot of our highly trained scientists time that they can use to focus on thinking about the scientific problem, not doing these routine tasks in the lab.”
Organa’s most sophisticated feature is perhaps its ability to provide feedback on generated data. “We were surprised to find that this visual language model can spot outliers on chemistry graphs,” explains Shkurti. The system also flags these ambiguities or uncertainties and suggests methods of troubleshooting.
The group is now working on improving the LLM’s ability to plan tasks and then revise those plans to make the system more amenable to experimental uncertainties.
“There’s a lot roboticists have to offer to scientists in order to amplify what they can do and get them better data,” Shkurti says. “I am really excited to try to create new possibilities.”
Kristel Tjandra is a freelance science writer based in Oahu.