AM-IA brings together news, research papers and applications relating to artificial intelligence published around the world into a single feed — and makes them available in your European language at the click of a button.
Scroll to browse · Click to filter
▶
Eugene Cheah (Featherless AI) détaille quatre étapes pour faire baisser la facture d'inférence : mesurer, dimensionner au plus juste, bien acheter sa capacité et revoir ses processus.
Researchers from Niklas Boers’ team reduce both the spatial and temporal scales of temperature and precipitation without having to retrain the model for each task.
By simulating the pulsations of rotating δ Scuti stars, two astronomers have determined the age of the cluster from five stars observed by TESS, in line with conventional methods.
The Japanese NAND memory manufacturer, once Japan’s largest company by market capitalisation, is suffering from rising interest rates, which are putting pressure on data centres, and from increasing competition from Chinese rivals.
Tomáš Čupr has staff interviewed by agents to map out undocumented processes, and then automates them. Balderton leads the Series A funding round and Deloitte becomes a partner.
The ‘ambient’ assistant listens in on conversations to learn its user’s habits and initiate actions without waiting for instructions. A dedicated connected device is planned.
The former ConductorOne, a specialist in identity governance, is set to make Workday the central hub for managing the access rights granted to staff and agents within the organisation. The price has not been disclosed.
The start-up promises to automatically close attack vectors by configuring the security tools already in place. In-Q-Tel and Microsoft’s fund are participating in this Series B round.
The London-based notified body, which specialises in medical software and AI, aims to reduce certification lead times, which currently range from 12 to 18 months. Headline is leading this Series B funding round.
Speaking on stage at TechCrunch Disrupt, Amjad Masad explained that Replit had switched to OpenAI for reasons of cost and performance. Sources close to the matter confirm that the partnership is still in place.
▶
Chez Apollo, construire à la main un portefeuille type de CLO prenait deux à trois jours ; des agents IA le font désormais en quelques minutes. Entretien filmé à AI Engineer New York 2026.
By representing the question bank as a knowledge graph, MARS streamlines level updates and maximises scores under time constraints, reducing time-overruns from 50 per cent to 5.6 per cent.
The method converts first-person RGB video footage into vision-language-action training data, achieving a 45-point improvement in performance on tasks not covered by robotic demonstrations.
A new method applied to WASP-94A b utilises the temporal signature of limb asymmetry, reducing the error in the measured molecular abundances by a factor of 30 to 50.
With 40 realistic tasks and 56 attack scenarios, INFUSE achieves an average success rate of 19.6 per cent, which is reduced from 58.2 per cent to 14.3 per cent when the constraints are tightened.
An audit of 2,550 scripted conversations shows that all chatbots direct users to a helpline, but that their persistence in the face of refusals ranges from almost none to nearly ten follow-up attempts.
Out of 2,099 catchment areas tested between 2021 and 2023, this open probabilistic model provides more accurate predictions of floods with return periods of 1.5 and 2 years than the AIFL method on which it is based.
Following a year of scandals, Apple is revising its App Review guidelines: mandatory safeguards for user-generated content, an explicit ban on obfuscated code used to conceal functions, and tougher penalties.
An internal report by Anthropic quantifies the impact of its own agents: the proportion of junior technical recruits has fallen from 27 per cent to 9 per cent, productivity has risen, and human attention has become the bottleneck.
A survey carried out in August 2026 amongst a representative sample of American adults and Anthropic staff reveals agreement on what AI is capable of today, but a wide divergence of views on what lies ahead.