Mathematics is often regarded as the ideal domain for measuring AI progress effectively. Math’s step-by-step logic is easy to track, and its definitive automatically verifiable answers remove any human or subjective factors. But AI systems are improving at such a pace that math benchmarks are struggling to keep up.

Way back in November 2024, non-profit research organization Epoch AI quietly released FrontierMath. A standardized, rigorous benchmark, Frontier Math was designed to measure the mathematical reasoning capabilities of the latest AI tools.

“It’s a bunch of really hard math problems,” explains Greg Burnham, Epoch AI Senior Researcher. “Originally, it was 300 problems that we now call tiers 1–3, but having seen AI capabilities really speed up, there was a feeling that we had to run to stay ahead, so now there’s a special challenge set of extra carefully constructed problems that we call tier 4.”

To a rough approximation, tiers 1–4 go from advanced undergraduate through to early postdoc level mathematics. When introduced, state-of-the-art AI models were unable to solve more than 2% of the problems FrontierMath contained. Fast forward to today and the best publicly available AI models, such as GPT-5.2 and Claude Opus 4.6, are solving over 40% of FrontierMath’s 300 tiers 1–3 problems, and over 30% of the 50 tier 4 problems.

AI takes on PhD level mathematics

And this dizzying pace of advancement is showing no signs of abating. For example, just recently Google DeepMind announced that Aletheia, an experimental AI system derived from Gemini Deep Think, achieved publishable PhD level research results. Though obscure mathematically—calculating certain structure constants in arithmetic geometry called eigenweights—the result is significant in terms of AI development.

“They’re claiming it was essentially autonomous, meaning a human wasn’t guiding the work, and it’s publishable,” Burnham says. “It’s definitely at the lower end of the spectrum of work that would get a mathematician excited, but it’s new—it’s something we truly haven’t really seen before.”

To place this achievement in context, every FrontierMath problem has a known answer that a human has derived. Though a human could probably have achieved Aletheia’s result “if they sat down and steeled themselves for a week,” says Burnham, no human had ever done so.

Aletheia’s results and other recent achievements by AI mathematicians point to new, tougher benchmarks being needed to understand AI capabilities, and fast, because existing ones will soon become irrelevant. “There are easier math benchmarks that are already obsolete, several generations of them,” says Burnham. “FrontierMath will probably saturate [meaning state-of-the-art AI models score 100%] within the next two years; could be faster.”

The First Proof challenge

To begin to address this problem, on February 6, a group of 11 highly distinguished mathematicians proposed the First Proof challenge, a set of 10 extremely difficult math questions which arose naturally in the authors’ research processes, and whose proofs are roughly five pages or less and had not been shared with anyone. The First Proof challenge was a preliminary effort to assess the capabilities of AI systems in solving research-level math questions on their own.

Generating serious buzz in the math community, professional and amateur mathematicians, and teams including OpenAI, all stepped up to the challenge. But by the time the authors posted the proofs on February 14, no one had submitted correct solutions to all 10 problems.

In fact, far from it. The authors themselves only solved two of the 10 problems using Gemini 3.0 Deep Think and ChatGPT 5.2 Pro. And most outside submissions fared little better, apart from OpenAI and a small Aletheia team at Google DeepMind. With “limited human supervision” OpenAI’s most advanced internal AI system solved five of the 10 problems, with Aletheia achieving similar outcomes—results met with a spectrum of emotions by different members of the mathematics community, from awe to disappointment. The team behind First Proof plans an even tougher second round on March 14.

A new frontier for AI

“I think First Proof is terrific: it’s as close as you could realistically get to putting an AI system in the shoes of a mathematician,” says Burnham. Though he admires how First Proof tests AI’s mathematical utility for a wide range of mathematics and mathematicians, Epoch AI has its own new approach to testing—FrontierMath: Open Problems. Uniquely, the pilot benchmark consists of 16 open problems (with more to follow) from research mathematics that professional mathematicians have tried and failed to solve. Since Open Problems’ release on January 27, none have been solved by an AI.

“With Open Problems, we’ve tried to make it more challenging,” says Burnham. “The baseline on its own would be publishable, at least in a specialty journal.” What’s more, each question is designed so that it can be automatically graded. “This is a bit counterintuitive,” Burnham adds. “No one knows the answers, but we have a computer program that will be able to judge whether the answer is right or not.”

Burnham sees First Proof and Open Problems as being complementary. “I would say understanding AI capabilities is a more-the-merrier situation,” he adds. “AI has gotten to the point where it’s, in some ways, better than most PhD students, so we need to pose problems where the answer would be at least moderately interesting to some human mathematicians, not because AI was doing it, but because it’s mathematics that human mathematicians care about.”

Survivor 50 is about to play the biggest game in Fiji history, and reality TV fans don’t have to break the bank to catch this season’s murderers’ row of castaways. Here’s the play:

You can watch Survivor for free in Canada. Traveling abroad in Australia, UK, Asia, India, etc? Theres a handy fix below (TLDR: get this VPN now to watch free anywhere).

The new skilled roles will be created as part of an initial five-year $350m investment.

AI infrastructure provider Equinix will create 200 jobs in Dundalk, Co Louth through an investment of up to $700m in a new facility that will be built by local company Hanley Energy.

New roles at the facility will be in technical fields such as precision engineering, quality assurance and lean manufacturing. Hiring has already begun, and there are plans for the facility to engage locally with apprenticeship and training programmes.

The deal guarantees an initial $350m across five years and is extendable to 10 years. The new facility will be used exclusively to build specialised power equipment for Equinix’s data centres worldwide.

Adaire Fox-Martin, the CEO and president of Equinix, said: “This investment builds upon Equinix’s longtime presence in Ireland and reflects the strategically important role the country plays in the global technology ecosystem.

“Our expansion in Dundalk further strengthens our ability to meet growing customer demand while creating local jobs and supporting the community.”

Equinix said that by centralising its production of components such as low-voltage switchgear, power distribution units and remote power panels at the new 150,000 sq ft facility, it expects to achieve up to 15pc faster lead times when compared to traditional procurement methods.

Taoiseach Micheál Martin, TD said: “This significant announcement reinforces Ireland’s position as a leader in digital infrastructure and advanced manufacturing. The creation of hundreds of skilled jobs and the introduction of world-class facilities in Dundalk is a major boost for the region and for our national economy.”

Equinix, which runs 280 data centres across six continents, works with partners such as Amazon Web Services, Nvidia, Azure and Google Cloud.

Managing director of Equinix Ireland Peter Lantry said the investment would secure the company’s supply chain locally in the Louth region and ensure its long-term presence there, while IDA Ireland CEO Michael Lohan said it would “deliver significant economic benefits and high-value jobs in Dundalk and the wider region”.

Hanley Energy, which employs around 850 worldwide, plans to build the new facility using “low-carbon materials and efficient construction practices”, according to Equinix. The facility will also feature a temperature-controlled testing laboratory – which Equinix claimed will be the only one of its kind in Ireland or the UK.

Hanley CEO John O’Driscoll said: “Partnering with Equinix on this transformative project highlights the strength of Irish engineering and innovation. Our advanced testing facilities and expertise will ensure that the equipment produced here meets the highest global standards, supporting data centres worldwide.”

At the end of last year, Hanley was acquired by US engineering and manufacturing company Jabil for around $725m.

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

Riley Walz, a software engineer famous for his online stunts, is joining OpenAI to research and develop new ways for humans to interact with AI, WIRED has learned. An OpenAI spokesperson confirmed the hire.

Walz built a reputation as Silicon Valley’s jester and has created a series of viral web projects that double as social commentary. His most recent initiative, Jmail, lets users search Jeffrey Epstein’s emails as if they’re accessing his personal Gmail inbox. Another project, Find My Parking Cops, used publicly available data to reverse engineer San Francisco’s parking ticket system to show people exactly where each parking enforcement officer last wrote a ticket.

Now, Walz’s skills creating novel web experiences will be put to use in OAI Labs, a relatively new team led by research leader Joanne Jang. The team is secretive about what it’s been working on but has been tasked with “inventing and prototyping new interfaces for how people collaborate with AI,” according to Jang.

OpenAI has spent the past several years racing with Google and Anthropic to create new, compelling ways for people to use its AI models. While ChatGPT has been a hit with consumers, now reaching more than 800 million people every week, the company is eyeing new interfaces to improve these experiences. The move comes as millions of developers have started using coding agents such as Claude Code as their main interface to access AI models. With hires like Walz, OpenAI hopes to get ahead of the next big AI product.

Walz’s online stunts have landed him in hot water from time to time. The Find My Parking Cops website lasted just four hours before San Francisco city officials shut down the live data feed Walz’s tool relied on. A San Francisco Municipal Transportation Agency representative said at the time that it shut down the tool to ensure “employees are able to do their jobs safely and without disruption.”

It’s not always city officials giving him a hard time, though. After the CEO of UnitedHealthcare was shot dead in New York City, and police said the killer had fled on a CitiBike, Walz tried to analyze trip data he had previously scraped for a separate project to help with the search. Walz told The New York Times that people online called him a “bootlicker” for helping authorities and threatened his safety.

Full Circle, the developer behind the new Skate game, has announced that it is restructuring and laying off staff. It’s not yet clear how many roles will be impacted by the changes, but the restructuring is happening less than six months after skate. launched in early access on September 15, 2025.

“We’re reshaping Full Circle to better support skate.’s long-term future,” Full Circle says. “These shifts mean making changes to our team structure, and some roles will be impacted. The teammates affected are talented colleagues and friends who helped build the foundation of skate. Their creativity and dedication are deeply ingrained in what players experience today. This decision is not a reflection of their impact and we’re committed to supporting them through this transition.”

Engadget has contacted Full Circle’s owner EA for more information about the layoffs. We’ll update this article if we hear back.

EA originally formed Full Circle in 2021 with a staff of development talent from the original Skate team. Skate was often positioned as a more realistic competitor to the Tony Hawk’s Pro Skater series, but the new studio has ultimately taken the franchise in a slightly different direction than fans may have expected. Previous Skate games were paid experiences with single-player and multiplayer modes, while skate. is a free-to-play live-service game supported with microtransactions.

Recent history, both the failure of Concord and the ongoing struggles of Highguard, serves as a testament to how hard it is to launch a live service game in the 2020s. Full Circle’s announcement notes the “tens of millions” of players that have tried the new game, but it’s possible a struggle to keep players interested and spending on microtransactions could be why it’s restructuring.