Diffusion Large Language Models (dLLMs) break the rigid left-to-right constraint of traditional LLMs, enabling token generation in arbitrary orders. Intuitively, this flexibility implies a solution space that strictly supersets the fixed autoregressive trajectory, theoretically unlocking superior reasoning potential. Indeed, for specific constraint satisfaction tasks (e.g., sudoku puzzles), this capability has proven to be highly advantageous. However, in this paper, we reveal that for general reasoning tasks (e.g., mathematics and coding), arbitrary order generation may in fact limit the reasoning potential of dLLMs. We find that dLLMs tend to exploit this order flexibility to bypass high-uncertainty tokens that are crucial for exploration, leading to a premature collapse of solution coverage. This observation motivates a rethink of RL approaches for dLLMs, where considerable complexities, such as handling combinatorial trajectories and intractable likelihoods, are often devoted to preserving this flexibility. We demonstrate that effective reasoning can be better elicited by simply forgoing arbitrary order and applying standard Group Relative Policy Optimization (GRPO) instead. Our approach, JustGRPO, is minimalist yet surprisingly effective (e.g., 89.1% accuracy on GSM8K) while fully retaining the parallel decoding ability of dLLMs.

Angel Reese thanks WNBA for suspending Sandy Brondello, frames'protected species' comment as discrimination Rory McIlroy should be the last golfer to criticize others for'performative' behavior Brexton Busch wins first race since father Kyle Busch's death in emotional return to Victory Lane Caitlin Clark'cheering hard' for Argentina over Spain in World Cup final: 'I want Messi' Ella Langley turns 63-year-old former NFL head coach into'Ella Fella' I am once again begging the folks in USC's athletic department to study some Greek literature Jaxson Dart's swimsuit model girlfriend gets patriotic, golfer Hailey Ostrom takes on Lake Powell & Bigfoot Caitlin Clark's former teammate calls out WNBA for suspending Tempo head coach over'protected species' remark Mark Levin: Trump is the ONLY president to recognize this... Iranian regime has proven they will not abide by'any agreement': Former Israeli ambassador Christopher Nolan brings'The Odyssey' to life with groundbreaking IMAX technology WATCH: Anchor STUNNED by socialist's answer More than 200 military musicians bring America's story to life The World Cup final on Sunday provided a tense, low-scoring finish to a memorable tournament held across the United States. Spain beat Argentina 1-0 in extra time to win the World Cup, but President Donald Trump ended up sharing plenty of the spotlight after joining the trophy ceremony. Rodri of Spain is presented with the FIFA World Cup Winner's Trophy by FIFA President Gianni Infantino and President Donald Trump after Spain's victory over Argentina in the FIFA World Cup 2026 final at New York New Jersey Stadium in East Rutherford, N.J., on July 19, 2026. FIFA President Gianni Infantino and President Donald J. Trump present the FIFA World Cup trophy to Rodri of Spain during the FIFA World Cup 2026 final match between Spain and Argentina at New York New Jersey Stadium in East Rutherford, N.J. (Europa Press Sports/Europa Press) Trump, alongside FIFA president Gianni Infantino, walked to the stage, drawing a mix of cheers and boos as he greeted Spain's players before the trophy presentation. Fox WORLD CUP ANNOUNCER SAYS TRUMP'S APPEARANCE AT FINAL IS'FUN, UNIQUE THING' DESPITE POTENTIAL BACKLASH Trump was briefly part of Spain's team photo as the players hoisted the trophy before Infantino stepped in and suggested he move aside.

The forty-third International Conference on Machine Learning (ICML) took place in Seoul, South Korea from 6-11 July. We take a look at what the participants got up to during the event. The PC Chairs are presenting the welcome remarks. One of the best conference dinners I've ever had. Try TimeChat-Captioner (at #ICML2026) -- a videoLLM that generates dense, time-aware captions for long videos.

On the 31st March, our editorial team headed to the Royal Society for AI for Science . This day-long conference explored how AI is changing the nature of scientific discovery, and was hosted by the Alan Turing Institute. The recordings from the event are now available on YouTube and are well worth a watch. You can read Ella Scallan's blog post about the day here . Lucy Smith is Senior Managing Editor for AIhub.

The Future of AI Research report, published in March 2025, aims to clearly identify the trajectory of AI research in a structured way. The report was led by outgoing AAAI President Francesca Rossi and covers 17 different AI topics . Members of the report team, and other selected AI practitioners, are taking part in a series of video panel discussions covering selected chapters from the report. In the sixth discussion in the collection, the three panellists tackle factuality and trustworthiness. Understanding factuality: why preventing false outputs from large language models remains AI's toughest problem Lucy Smith is Senior Managing Editor for AIhub.

People often make decisions through "satisficing," gathering just enough information to make a satisfactory prediction of a likely outcome. A series of experimental games shows that people also employ satisficing to learn social rules and conventions. This finding offers new insight into social learning and reveals a key difference between how humans and LLMs make predictions. The premise of AI large language models is that any problem can be solved by vacuuming up as much information as possible, running it through probability models, and performing complex calculations to make predictions and come up with the optimal solution. Another premise behind LLMs is that they emulate the way human brains operate.

Things to Do in L.A. Tap to enable a layout that focuses on the article. Omar Yaghi, professor at the University of California, Berkeley, speaks during a media conference in Brussels, Oct. 8, 2025, after being one of three scientists awarded the Nobel Prize in chemistry. This is read by an automated voice. Please report any issues or inconsistencies here . See more from the L.A. Times in Google Search.

This year, RoboCup took place in Incheon, South Korea, from 2-6 July. The event saw teams take part in competitions, training sessions, and a symposium. Take a look at what the participants got up to in our round up from social media. RoboCup 2026 officially begins today! A post shared by RoboCup Federation (@robocup.official)

The EurAI Distinguished Service Award started in 2012, and it is presented annually to individuals who have made exceptional contributions to the European AI community. This year, the award goes to two researchers: Jérôme Lang and Luc de Raedt. Find out who won the small, middle and large divisions in Incheon. Find out the latest from day two of the competition. In the first of our round-ups from the humanoid league we introduce the competition, and report some preliminary results.

Jamir Nazir, the controversial winner of the Commonwealth award, tells his side of the story. Jamir Nazir has become the face of the AI-writing crisis. In May, the largely unknown 62-year-old Trinidadian writer was named a regional winner of the prestigious Commonwealth Prize for his short story " The Serpent in the Grove " But after it was published in the literary magazine, signs began to emerge that the story--about a cocoa farmer who cheated on his wife, and then tried to kill her--may have been AI-generated. Inscrutable lines plucked from Nazir's dense prose were mocked and memed. A young woman in the story "had the kind of walking that made benches become men."