Technology

OpenAI and Anthropic claim to have disproved the exponential time hypothesis, a pillar of theoretical computer science
OpenAI and Anthropic said this week they have disproved the exponential time hypothesis (ETH), an assumption underpinning much research in complexity and algorithm design. Both companies published the proofs on GitHub, in natural language and in machine-checkable Lean format. Researchers have not yet independently verified them and warn that the result, if confirmed, would render numerous papers obsolete, including those of academics presenting at the IEEE Symposium on Foundations of Computer Science.
TechnologyInternationalWIREDOct 29AI
Iryo CEO warned in the Senate about failures in the LZB safety system
Fabrizio Favara, CEO of Iryo, told the Senate committee investigating the state of Spain's rail network that the LZB signalling system is "safe, but quite old", and that whenever the train signal disconnects from the ground signal the convoy stops and takes time to resume service. The report also includes audio from an Iryo driver who could not restart the system before an accident on the Madrid-Andalusia corridor.
TechnologyInternationalEl CorreoOct 29AI
China unveils Sugon 8000, a supercluster with 100,000 domestically made GPUs
The Sugon 8000, a supercluster with around 100,000 graphics cards designed and manufactured in China, was presented at China's National High Performance Computing Conference in Shenzhen. The system is located in Zhengzhou, Henan province, and is the central node of China's National Supercomputing Internet. Sugon says all components —chips, networking, storage, cooling and software— come from Chinese suppliers. The company was added to the US Entity List in 2019.
TechnologyInternationalXatakaOct 29AI
How to spot a fake image made with artificial intelligence, according to three experts
Advances in artificial intelligence have reduced the visible errors in generated images, so looking for deformed hands or meaningless text is no longer enough. Three experts consulted by La Tercera recommend checking hands, faces, shadows, reflections and background elements, and observing how objects and people relate to each other. They note that visual inspection has limits: they cite a 2025 Microsoft Research study in which people were correct in about 62% of cases. They suggest using reverse image search and detection tools as clues, not as proof.
TechnologyInternationalLa TerceraOct 29AI
Microbiologist Mario Feldman warns "we are losing the battle against bacteria"
Mario Feldman, the first graduate in Biotechnology from the National University of Rosario and now a professor of Molecular Microbiology at Washington University in St. Louis, warns about bacterial resistance to antibiotics, a "silent pandemic." His laboratory studies Acinetobacter baumannii and strategies that target virulence factors instead of killing the microorganism. He also co-founded Omniose, which develops bacterial vaccines with a biological platform and collaborates with AstraZeneca, and received about 15 million dollars in research grants.
TechnologyInternationalLa NaciónOct 29AI
Wittmann expects €330 million in revenue, up 10%, and boosts local production and digitalization
The Wittmann Group sees a recovery in some markets but considers it premature to speak of a consolidated trend change, according to owner and CEO Michael Wittmann at the Fakuma 2026 press conference. For the current fiscal year it expects revenue of around €330 million, 10% more than the previous year. The group employs 2,050 people and has ten production plants and 35 subsidiaries. The company remains cautious due to the transformation of injection molding in Europe and the automotive sector.
TechnologyInternationalInterempresas MediaOct 29AI
Alphabet plunges on Wall Street and Microsoft rebounds after earnings: the uneven impact of AI
Alphabet fell more than 10% in after-hours trading, a loss of nearly $400 billion in market capitalization, after missing forecasts and posting negative free cash flow again due to its AI investments. Microsoft, by contrast, rebounded 4% after beating estimates, with Azure growing 48%. Alphabet invested $51.3 billion between July and September and spent nothing on share buybacks.
TechnologyInternationalCinco DíasOct 29AI
PROhumana warns that AI is moving faster than Chilean corporate boards
PROhumana's Round Table, which brought together 46 leaders from 38 companies and organizations in Chile, concluded that artificial intelligence is moving faster than boards, governance structures and worker preparation. Its founder, Soledad Teixidó, said general managers are facing the change while their boards are not moving at the same pace, and that companies are holding back new hiring although they are not yet cutting jobs.
TechnologyInternationalEl MostradorOct 29AI
Apple quietly cuts the student discount on the MacBook Neo
Apple reduced the student discount on the MacBook Neo minutes before announcing the touchscreen MacBook Pro. The laptop, launched in March from 599 euros for students and teachers, now costs 739 euros for that group in its base version, 23.37% more. In late June there was already a 100-euro increase across all Macs, including student prices. The gap with the general price is now 60 euros. The MacBook Air and MacBook Pro keep their student prices.
TechnologyInternationalApplesferaOct 29AI
China condemns US ban on three genetics companies over alleged ties to Uyghur forced labor
China's Ministry of Commerce condemned US restrictions on BGI Genomics, MGI Tech and Shenzhen BGI Technology (BGI Group), added to a list of entities barred from imports over alleged links to Uyghur forced labor in Xinjiang. Beijing called the measure protectionist and warned it will take the necessary steps to defend its interests. BGI Group said the restrictions will not have a significant impact on its operations and MGI Tech announced it will appeal its listing. The US list now totals 190 entities.
TechnologyInternationalInfobaeOct 29AI
SK Hynix lifts operating profit by 632% and warns memory shortage will last until 2030
SK Hynix posted the best quarter in its history: operating profit from July to September reached 83.33 trillion won (about 55.61 billion euros), up 632% year on year, and revenue topped 100 trillion won in a single quarter for the first time. Operating margin hit a record 79.7%. The company, the leader in HBM memory and a supplier to Nvidia, says its DRAM and HBM output is practically sold out through 2029 and expects the shortage to last until 2030.
TechnologyInternationalXatakaOct 29AI
China's Lao Gan Ma reports record 2025 revenue and faces reputational crisis over AI use
Chinese chili sauce brand Lao Gan Ma reported revenue of about 540 billion...
TechnologyInternationalXatakaOct 29AI
A solar storm in November 2025 caused GPS errors of more than 10 meters over the US
A study led by space physicist Endawoke Yizengaw of The Aerospace Corporation, published in Geophysical Research Letters, reconstructed a solar storm in November 2025 that disturbed the ionosphere over North America. Researchers detected a band of high electron density that crossed the ionosphere from east to west, with sharp changes at its edges that generated irregularities and amplitude scintillation in navigation signals. In some areas the horizontal GPS error exceeded 10 meters.
TechnologyInternationalEl ConfidencialOct 29AI
China achieves 270 °C industrial heat without fossil fuels using a thermoacoustic heat pump
Scientists at the Technical Institute of Physics and Chemistry of the Chinese Academy of Sciences built a heat pump that reaches 270 °C using sound waves and no moving mechanical components. According to results published in the journal Energy, the system recovers heat at about 145 °C and raises it to a higher temperature. In the experiments one unit required a heat source of about 300 degrees and a thermal power of 2.7 kilowatts was achieved, with efficiency falling as the output temperature rose.
TechnologyInternationalLa VanguardiaOct 29AI
The Spanish engineer leading the IOC's AI agenda: from cyberbullying of athletes to supporting judges
Alejandro Merino-Madrid, a 39-year-old aerospace engineer from Logroño, joined the International Olympic Committee in 2024 to lead its Olympic AI Agenda. The project includes a system that detects cyberbullying against athletes, an guidance app for athletes in the Olympic Village, tools that help judges spot objective infractions and tablet-based tests to scout young talent in Dakar. Of 2.4 million interactions examined at this year's Winter Games, 10,200 were hate messages.
TechnologyInternationalMARCAOct 29AI
ETUI researcher: Europe's AI law regulates products, not human relations
Aída Ponce Del Castillo, senior researcher at the European Trade Union Institute (ETUI), argues that Europe's artificial intelligence law regulates products rather than human relations, leaving worker and candidate protections incomplete. She warns that algorithmic management blurs the relationship between managers and employees and that the roles most exposed to automation are those of managers. She advocates for unions to negotiate software purchases before deployment and calls for rights to information, objection, explanation and review.
TechnologyInternationalValencia PlazaOct 29AI
Lazarus, the North Korean hacking group, uses AI and post-quantum encryption in its attacks
Kaspersky found that the Lazarus group, unofficially linked to North Korea, used post-quantum encryption in a cyberespionage campaign against South Korean construction, software and telecommunications companies. The attackers used AI to create fake identities and emails with links to fake résumés that downloaded spyware. Post-quantum encryption, designed to withstand quantum computers, makes it harder for defenders and researchers to recover the data. It is the first known use of this technique by Lazarus, after the ransomware case attributed to Kyber in March.
TechnologyInternationalEL PAÍSOct 28AI
Anthropic acknowledges intense AI competition and leaves door open to new acquisitions ahead of its IPO
In the prospectus for its initial public offering, expected in November, Anthropic warns of intense competition in the market for AI models, platforms and applications, with Google and OpenAI as its main rivals. Dario Amodei's company notes that many competitors have substantially greater financial, technical and marketing resources, and leaves the door open to selective acquisitions, investments and alliances to bring in technology and talent.
TechnologyInternationalCinco DíasOct 28AI
Spanish employees use AI twice as much as their bosses: 88% versus 46%
A study by consultancy EY indicates that in Spain 88% of employees have already incorporated artificial intelligence into their daily work, nearly double the share of leaders, among whom daily use reaches 46%. According to InfoJobs, 82% of employees use AI tools in their jobs. Experts attribute the gap to the nature of managerial roles and to a lack of training: 71% of managers have received no guidance on leading teams in an AI environment.
TechnologyInternationalEl ConfidencialOct 28AI
AI begins screening candidates at temporary work agencies in Burgos
Temporary work agencies in Burgos are progressively adopting artificial intelligence tools to speed up recruitment: they identify and match profiles to vacancies, manage applications and automate repetitive tasks. The systems use ATS filters that discard or select candidates based on location, education, experience or keywords, plus chatbots that ask applicants basic questionnaires. Adecco and Randstad stress that the technology does not replace professional judgement in hiring and that it helps reduce demographic bias.
TechnologyInternationalDiario de BurgosOct 28AI