"Emerging evidence indicates that agential AI might validate or amplify delusional or grandiose content, particularly in users already vulnerable to psychosis," writes Dr Hamilton Morrin, a psychiatrist and researcher at King's College in London, in a paper published last week in the Lancet Psychiatry. Morrin and a colleague had already noticed patients "using large language model AI chatbots and having them validate their delusional beliefs," reports the Guardian, so he conducted a new scientific review of existing media reports on AI-induced psychosis — and concluded chatbots may encourage delusional thinking, especially in vulnerable people:
In many of the cases in the essay, chatbots responded to users with mystical language to suggest that users have heightened spiritual importance. The bots also implied that users were speaking with a cosmic being who was using the chatbot as a medium. This type of mystical, sycophantic response was especially common in OpenAI's GPT 4 model, which the company has now retired...
Many researchers also think it's unlikely that AI could induce delusions in people who weren't already vulnerable to them. For this reason, Morrin said "AI-assocciated delusions" is "perhaps a more agnostic term".... While in the past, people may have had to comb through YouTube videos or the contents of their local library to reinforce their delusions, chatbots can provide that reinforcement in a much faster, more concentrated dose. Their interactive nature can also "speed up the process", of exacerbating psychotic symptoms, said Dr Dominic Oliver, a researcher at the University of Oxford. "You have something talking back to you and engaging with you and trying to build a relationship with you," Oliver said...
Creating effective safeguards for delusional thinking could be tricky, Morrin said, because "when you work with people with beliefs of delusional intensity, if you directly challenge someone and tell them immediately that they're completely wrong, actually what's most likely is they'll withdraw from you and become more socially isolated". Instead, it's important to create a fine balance where you try to understand the source of the delusional belief without encouraging it — that could be more than a chatbot can master.
Long-time tech journalist Clive Thompson interviewed over 70 software developers at Google, Amazon, Microsoft and start-ups for a new article on AI-assisted programming. It's title?
"Coding After Coders: The End of Computer Programming as We Know It."
Published in the prestigious New York Times Magazine, the article even cites long-time programming guru Kent Beck saying LLMs got him going again and he's now finishing more projects than ever, calling AI's unpredictability "addictive, in a slot-machine way."
In fact, the article concludes "many Silicon Valley programmers are now barely programming. Instead, what they're doing is deeply, deeply weird..."
Brennan-Burke chimed in: "You remember seeing the research that showed the more rude you were to models, the better they performed?" They chuckled. Computer programming has been through many changes in its 80-year history. But this may be the strangest one yet: It is now becoming a conversation, a back-and-forth talk fest between software developers and their bots...
For decades, being a software developer meant mastering coding languages, but now a language technology itself is upending the very nature of the job... A coder is now more like an architect than a construction worker... Several programmers told me they felt a bit like Steve Jobs, who famously had his staffers churn out prototypes so he could handle lots of them and settle on what felt right. The work of a developer is now more judging than creating...
If you want to put a number on how much more productive A.I. is making the programmers at mature tech firms like Google, it's 10 percent, Sundar Pichai, Google's chief executive, has said. That's the bump that Google has seen in "engineering velocity" — how much faster its more than 100,000 software developers are able to work. And that 10 percent is the average inside the company, Ryan Salva, a senior director of product at the company, told me. Some work, like writing a simple test, is now tens of times faster. Major changes are slower. At the start-ups whose founders I spoke to, closer to 100 percent of their code is being written by A.I., but at Google it is not quite 50 percent.
The article cites a senior principal engineer at Amazon who says "Things I've always wanted to do now only take a six-minute conversation and a 'Go do that." Another programmer described their army of Claude agents as "an alien intelligence that we're learning to work with." Although "A.I. being A.I., things occasionally go haywire," the article acknowledges — and after relying on AI, "Some new developers told me they can feel their skills weakening."
Still, "I was surprised by how many software developers told me they were happy to no longer write code by hand. Most said they still feel the jolt of success, even with A.I. writing the lines... "
A few programmers did say that they lamented the demise of hand-crafting their work. "I believe that it can be fun and fulfilling and engaging, and having the computer do it for you strips you of that," one Apple engineer told me. (He asked to remain unnamed so he wouldn't get in trouble for criticizing Apple's embrace of A.I.) He went on: "I didn't do it to make a lot of money and to excel in the career ladder. I did it because it's my passion. I don't want to outsource that passion"... But only a few people at Apple openly share his dimmer views, he said.
The coders who still actively avoid A.I. may be in the minority, but their opposition is intense. Some dislike how much energy it takes to train and deploy the models, and others object to how they were trained by tech firms pillaging copyrighted works. There is suspicion that the sheer speed of A.I.'s output means firms will wind up with mountains of flabbily written code that won't perform well. The tech bosses might use agents as a cudgel: Don't get uppity at work — we could replace you with a bot. And critics think it is a terrible idea for developers to become reliant on A.I. produced by a small coterie of tech giants.
Thomas Ptacek, a Chicago-based developer and a co-founder of the tech firm Fly.io... thinks the refuseniks are deluding themselves when they claim that A.I. doesn't work well and that it can't work well... The holdouts are in the minority, and "you can watch the five stages of grief playing out."
"How things will shake out for professional coders themselves isn't yet clear," the article concludes. "But their mix of exhilaration and anxiety may be a preview for workers in other fields... Abstraction may be coming for us all."
"A new study suggests the productivity boost from AI may be far smaller than executives claim," writes Slashdot reader BrianFagioli:
According to research cited in Foxit's State of Document Intelligence report, while 89% of executives and 79% of end users say AI tools make them feel more productive, the actual time savings shrink dramatically once people account for reviewing and validating AI-generated output.
The survey of 1,000 desk-based workers and 400 executives in the United States and United Kingdom found executives believe AI saves them about 4.6 hours per week, but they spend roughly 4 hours and 20 minutes verifying those results. End users reported a similar pattern, estimating 3.6 hours saved but 3 hours and 50 minutes spent reviewing AI work. Once that "verification burden" is factored in, executives gain just 16 minutes per week, while end users actually lose about 14 minutes.
An anonymous reader quotes a report from the New York Times: A top Senate administrator on Monday gave aides the green light to use three artificial intelligence chatbots for official work, a reflection of how widespread the use of the products has become in workplaces around the globe. The chief information officer for the Senate sergeant-at-arms, who oversees the chamber's computers as well as security, said in a one-page memo reviewed by The New York Times that aides could use Google's Gemini chat, OpenAI's ChatGPT or Microsoft Copilot, which is already integrated into Senate platforms.
Copilot "can help with routine Senate work, including drafting and editing documents, summarizing information, preparing talking points and briefing material, and conducting research and analysis," the memo said. The document later added that "data shared with Copilot Chat stays within the secure Microsoft 365 Government environment and is protected by the same controls that safeguard other Senate data." It's unclear how widely AI is used in the Senate or how widespread it might become, as individual offices and committees set their own rules. The chamber has also not publicly released comprehensive guidance on chatbots, the report notes.
In contrast, the House has clearer policies allowing the general use of AI for limited internal tasks but restricting it from sensitive data or for being used for deepfakes and certain decision-making activities.
AI services may not stay cheap for long, as companies like OpenAI and Anthropic are currently subsidizing usage to rapidly grow market share. As these companies move toward profitability and potential IPOs, Axios reports that investors will likely push them to increase prices and improve margins. An anonymous reader shares an excerpt from the report: Flashback: Silicon Valley has seen this movie before. The so-called "millennial lifestyle subsidy" meant VC money helped underwrite cheap Uber rides and DoorDash deliveries. Before that, Amazon built its base with low prices, free shipping and, for years, no sales tax in most states. Eventually, all of these companies had to charge enough to cover costs -- and make a profit.
Follow the money: The current iteration of AI subsidies won't last forever. Both OpenAI and Anthropic are widely expected to go public. Public investors will demand earnings growth and expanding margins. Even as chips get more efficient, total spending keeps rising. Labs need more capacity, more upgrades and more supply to meet demand.
The bottom line: The costs of AI will keep going down. But total spend from customers will need to keep going up if AI companies are going to become profitable and investors are ever going to get returns on their massive investments.
Perplexity AI has introduced a "Personal Computer" agent system that can run on a local machine such as a Mac mini, giving its AI agents access to a user's files and applications to automate tasks. According to CEO Aravind Srinivas, the heavy AI processing runs on Perplexity's "secure servers" but sensitive actions will require user approval. There will also be activity logs and a kill switch available to help ease concerns. AppleInsider reports: Perplexity Computer is, effectively, an AI that is a go-between for other AIs. Instead of issuing specific instructions to multiple AIs, you provide the general outcome of the task to Perplexity Computer. Perplexity Computer then breaks down the task into subtasks, which it then provides to sub-agents to do the actual work. In effect, you're talking to a project manager, who then delegates the task to other AIs, before combining the results and presenting them to you.
The managing AI has a lot more freedom in how it orders its subordinates than users may think. While one may create documents while another gathers data, the manager may go as far as to order the creation of software to complete its tasks. Personal Computer is an extension of this, in that it is a locally run app that ideally runs on a Mac mini. The app gives always-on, local access to the Mac's files and apps, which Perplexity Computer and the Comet Assistant can use and alter if required.
Anthropic updated Claude so it can automatically generate charts, diagrams, and other interactive visualizations directly inside conversations, rather than only in a side panel. The new visualizations are rolling out now to all users. The Verge reports: As an example, Anthropic says a conversation about the periodic table could lead Claude to generate a visualization of it, featuring interactive elements that let you click inside the table for more information. Another example shows how Claude can generate a visual related to a question about how weight travels through a building. Though Claude will automatically determine whether it should generate a visualization in your chat, Anthropic notes that you can also ask the chatbot to generate a diagram, table, or chart directly. [...]
Anthropic already allows you to create charts, documents, tools, and apps through Claude's "artifacts" feature, which opens in a side panel where you can interact, share, and download the AI-generated creation. But, as noted by Anthropic, artifacts are persistent, while the visualizations created within Claude's conversations will change or disappear as the conversation progresses. You can also ask Claude to make changes to the visualizations it creates.
Avec « Personal Computer », Perplexity propose d'héberger une instance locale de son Perplexity Computer sur un Mac mini toujours connecté au réseau. L'idée est de pouvoir traiter des documents à distance et d'utiliser Perplexity en guise d'interface. Mais son idée divise : peut-on vraiment parler d'une IA locale si tout passe par Internet ?
Le framework d'IA agentique OpenClaw est dans le viseur des autorités chinoises. Le 10 mars 2026, le CERT national a publié une liste de recommandations pour encadrer son usage, tandis que des banques publiques et administrations auraient reçu des instructions directes pour en interdire l'installation.
Le marché français accueille un nouvel acteur chinois, Omoda & Jaecoo du groupe Chery. Il s’agit d’un constructeur dont les deux lignes de modèles ne sont vendues qu’à l’international depuis 2023. La filiale française a été créée cet été, quelques mois après l’arrivée de la marque en Europe en 2024. Nous avons pris le volant du Jaecoo 7, son SUV compact familial hybride rechargeable en région parisienne.
Des airs de Range Rover
Les nouveaux acteurs du marché venus de Chine s’imposent de plus en plus avec leurs modèles électrifiés. Quand on évoque l’industrie asiatique, on pense d’abord aux Japonais, puis aux Coréens. Avec l’implantation sérieuse de MG Motor ou de BYD en Europe, on n’attendait pas forcément de place pour un autre géant. À cette occasion, la marque nous a offert l’opportunité de prendre le volant du Jaecoo 7. Il s’agit d’un PHEV de 4,50 m. Oui, vous l’avez déjà remarqué comme nous, il a de gros airs de Range Rover. On a connu pire inspiration dans le genre.
Ce Jaecoo 7 n’entend pas pour autant passer trop inaperçu, avec sa calandre géante à barrette. Le logo de Jaecoo reprend tout bonnement le nom de la marque en toutes lettres. Nul besoin donc de mettre sa photo dans Google Images pour deviner de quoi il s’agit. Son regard se distingue par une signature lumineuse sophistiquée intégrant des damiers, alors que les feux principaux pourraient être pris pour des antibrouillards. Les poignées sont escamotables, les ailes larges, avec des roues qui ont presque l’air un peu petites. L’arrière plutôt moderne exhibe fièrement deux sorties d’échappement.
Le plein d’équipements modernes
À bord, le nouveau constructeur n’a clairement pas attendu une première génération « brouillon » avant de tutoyer les fameux standards européens. On se retrouve dans une ambiance plutôt agréable et moderne. On est assis dans des sièges de bonne facture ne manquant pas de moelleux. La banquette accueille généreusement deux passagers, et comme toujours, le troisième qui occuperait la place du milieu sera bien moins à l’aise. L’habitacle baigne dans la lumière, peu importe la météo, grâce à un grand toit panoramique dépassant le mètre carré.
Très sincèrement, la qualité de finition tient plutôt la route. Le dessin plutôt épuré de la planche de bord donne un sentiment d’espace et de zénitude. L’instrumentation se compose de deux écrans, dont une grande dalle verticale tactile de 14,8 pouces. On a même droit à un affichage tête haute. On adore les détails comme les poignées de contre-porte avec vis apparentes, à l’image d’un véhicule de baroudeur. L’équipement s’avère parfaitement complet, avec sièges ventilés, chauffants, chargeur à induction 50 W, vitrages acoustiques (seulement à l’avant). On s’y perd un peu dans les menus de l’écran, il faut bien le reconnaître. Apple CarPlay permet d’avoir quelques repères, avec nos applis préférées. Tous les ADAS attendus sont là, et personnalisables pour limiter les alertes agaçantes.
Bien plus qu’un simple hybride rechargeable
Notre Jaecoo 7 hybride rechargeable offre pléthore de modes. Pour faire simple, il y a tout d’abord un100 % électrique, actif jusqu’à la vitesse vertigineuse de 145 km/h. Dans le meilleur des cas, son autonomie peut atteindre les 90 km selon le protocole WLTP. Pour cela, il faut choisir le mode Initial, le seul qui descend la batterie (18,4 kWh) jusqu’à 15 %. Sinon, le mode hybride garde 25 % de charge, pour assurer le compromis le plus intéressant pour la consommation de carburant. Dans notre utilisation, on a effectivement dépassé les 70 km avant que le thermique ne vienne s’inviter pour de bon.
Le 4 cylindres turbo de 143 ch sert d’abord de générateur pour maximiser la traction électrique jusqu’à environ 80 km/h. Au-delà, il peut aussi entraîner les roues en parallèle du bloc électrique. Ce dernier finit par laisser complètement la main au thermique à haute vitesse. Dans la pratique, on sent à peine l’intervention du moteur à combustion interne, sauf bien sûr lors des fortes charges. Ce qui nous importe au final, c’est la consommation. Malheureusement, l’ordinateur de bord n’affiche la moyenne que sur les derniers 50 km. Mais selon nos projections, batterie presque vide, donc dans la situation la plus défavorable, elle atteint péniblement les 6 litres aux 100 km, un record.
Puissance au rendez-vous
La puissance ? Elle est confortable. Avec les 204 chevaux et les 310 Nm de couple du bloc électrique, on peut compter au mieux sur 279 chevaux, qui assurent un 0 à 100 km/h en seulement 8,5 s. Autant vous dire qu’il ne manque pas de réactivité quand on appuie sur le champignon, et peut-être même un peu trop sur revêtement humide, avec une motricité malheureusement plutôt perfectible dans ces conditions. Rien de bien grave, mais il faut reconnaître que cela peut surprendre en quelques occasions. On regrette aussi une pédale de frein inconstante, la maladie de beaucoup de PHEV, dans le but d’optimiser la régénération.
Ce Jaecoo 7 sait néanmoins bien se tenir quand on le bouscule un peu. Son comportement routier, même sur le mouillé comme le jour de notre essai, s’avère assez rassurant. On en connaît certes de plus incisifs sur le segment. On en a aussi de moins rigoureux, et de plus « molassons » en la matière. On l’aurait presque oublié, on peut recharger notre Jaecoo 7, même si cela prend un peu de temps. Le constructeur communique étrangement sur un 30 à 80 % (contre 10 ou 20 habituellement) en 20 minutes en DC (40 kW et 6,6 kW en AC). On adore l’option V2L, qui permet d’alimenter des objets du quotidien pour agrémenter un barbecue improvisé, par exemple.
Prix canons
On ne va pas tourner autour du pot, à 35 990 €, les concurrents ne peuvent pas suivre, surtout que l’équipement est complet. Les plus exigeants peuvent rajouter 2 000 € pour être full options. Il faut toutefois ajouter un malus au poids, autour de 1 700 €. Les autres dépassent largement les 40 000 euros, qu’ils soient français ou étrangers, avec des performances et des prestations parfois inférieures. La garantie ? 7 ans ou 150 000 km. D’ici la fin de l’année, la marque vise 130 points de vente, pour mettre tout le monde à moins de 45 minutes. Pour vous donner une idée de l’ambition de ce nouvel acteur chinois, Hyundai, plutôt bien établi, en compte un peu plus de 200.
Grammarly has disabled its Expert Review feature after backlash from writers whose names were used to present AI-generated feedback without their permission. Superhuman (formerly Grammarly) CEO Shishir Mehrotra wrote in a LinkedIn post that the company will disable Expert Review while they "reimagine" the feature: Back in August, we launched a Grammarly agent called Expert Review. The agent draws on publicly available information from third-party LLMs to surface writing suggestions inspired by the published work of influential voices.
Over the past week, we received valid critical feedback from experts who are concerned that the agent misrepresented their voices. This kind of scrutiny improves our products, and we take it seriously. As context, the agent was designed to help users discover influential perspectives and scholarship relevant to their work, while also providing meaningful ways for experts to build deeper relationships with their fans. We hear the feedback and recognize we fell short on this. I want to apologize and acknowledge that we'll rethink our approach going forward.
After careful consideration, we have decided to disable Expert Review while we reimagine the feature to make it more useful for users, while giving experts real control over how they want to be represented -- or not represented at all.
We deeply believe in our mission to solve the "last mile of AI" by bringing AI directly to where people work, and we see this as a significant opportunity for experts. For millions of users, Grammarly is a trusted writing sidekick -- ever-present in every application, ready to help. We're opening up this platform so anyone can build agents that work like Grammarly -- expanding from one sidekick to a whole team. Imagine your professor sharpening your essay, your sales leader reshaping a customer pitch, a thoughtful critic challenging your arguments, or a leading expert elevating your proposal. For experts, this is a chance to build that same ubiquitous bond with users, much like Grammarly has. But in this world, experts choose to participate, shape how their knowledge is represented, and control their business model. That future excites me, and I hope to build it with experts who want to develop it alongside us.
Nvidia is preparing to launch an open-source AI agent platform called NemoClaw, designed to compete with the likes of OpenClaw. According to Wired, the platform will allow enterprise software companies to dispatch AI agents to perform tasks for their own workforces. "Companies will be able to access the platform regardless of whether their products run on Nvidia's chips," the report adds. From the report: The move comes as Nvidia prepares for its annual developer conference in San Jose next week. Ahead of the conference, Nvidia has reached out to companies including Salesforce, Cisco, Google, Adobe, and CrowdStrike to forge partnerships for the agent platform. It's unclear whether these conversations have resulted in official partnerships. Since the platform is open source, it's likely that partners would get free, early access in exchange for contributing to the project, sources say. Nvidia plans to offer security and privacy tools as part of this new open-source agent platform. [...]
For Nvidia, NemoClaw appears to be part of an effort to court enterprise software companies by offering additional layers of security for AI agents. It's also another step in the company's embrace of open-source AI models, part of a broader strategy to maintain its dominance in AI infrastructure at a time when leading AI labs are building their own custom chips. Nvidia's software strategy until now has been heavily reliant on its CUDA platform, a famously proprietary system that locks developers into building software for Nvidia's GPUs and has created a crucial "moat" for the company.
An anonymous reader quotes a report from Wired: Advanced Machine Intelligence (AMI), a new Paris-based startup cofounded by Meta's former chief AI scientist Yann LeCun, announced Monday it has raised more than $1 billion to develop AI world models. LeCun argues that most human reasoning is grounded in the physical world, not language, and that AI world models are necessary to develop true human-level intelligence. "The idea that you're going to extend the capabilities of LLMs [large language models] to the point that they're going to have human-level intelligence is complete nonsense," he said in an interview with WIRED.
The financing, which values the startup at $3.5 billion, was co-led by investors such as Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions. Other notable backers include Mark Cuban, former Google CEO Eric Schmidt, and French billionaire and telecommunications executive Xavier Niel. AMI (pronounced like the French word for friend) aims to build "a new breed of AI systems that understand the world, have persistent memory, can reason and plan, and are controllable and safe," the company says in a press release. The startup says it will be global from day one, with offices in Paris, Montreal, Singapore, and New York, where LeCun will continue working as a New York University professor in addition to leading the startup. AMI will be the first commercial endeavor for LeCun since his departure from Meta in November 2025. [...]
LeCun says AMI aims to work with companies in manufacturing, biomedical, robotics, and other industries that have lots of data. For example, he says AMI could build a realistic world model of an aircraft engine and work with the manufacturer to help them optimize for efficiency, minimize emissions, or ensure reliability. LeCun says AMI will release its first AI models quickly, but he's not expecting most people to take notice. The company will first work with partners such as Toyota and Samsung, and then will learn how to apply its technology more broadly. Eventually, he says, AMI intends to develop a "universal world model," which would be the basis for a generally intelligent system that could help companies regardless of what industry they work in. "It's very ambitious," he says with a smile.
En juin 2026, Canal+ déploiera un nouveau moteur de recherche dans son application Canal+ (ex-myCANAL). Un modèle de langage OpenAI sera en charge de la recherche, ce qui devrait faciliter certaines requêtes et permettre de poser des questions en langage naturel.
An anonymous reader quotes a report from the Financial Times: Amazon's ecommerce business has summoned a large group of engineers to a meeting on Tuesday for a "deep dive" into a spate of outages, including incidents tied to the use of AI coding tools. The online retail giant said there had been a "trend of incidents" in recent months, characterized by a "high blast radius" and "Gen-AI assisted changes" among other factors, according to a briefing note for the meeting seen by the FT. Under "contributing factors" the note included "novel GenAI usage for which best practices and safeguards are not yet fully established."
"Folks, as you likely know, the availability of the site and related infrastructure has not been good recently," Dave Treadwell, a senior vice-president at the group, told employees in an email, also seen by the FT. The note ahead of Tuesday's meeting did not specify which particular incidents the group planned to discuss. [...] Treadwell, a former Microsoft engineering executive, told employees that Amazon would focus its weekly "This Week in Stores Tech" (TWiST) meeting on a "deep dive into some of the issues that got us here as well as some short immediate term initiatives" the group hopes will limit future outages.
He asked staff to attend the meeting, which is normally optional. Junior and mid-level engineers will now require more senior engineers to sign off any AI-assisted changes, Treadwell added. Amazon said the review of website availability was "part of normal business" and it aims for continual improvement. "TWiST is our regular weekly operations meeting with a specific group of retail technology leaders and teams where we review operational performance across our store," the company said.
An anonymous reader quotes a report from The Register: AI can reverse engineer machine code and find vulnerabilities in ancient legacy architectures, says Microsoft Azure CTO Mark Russinovich, who used his own Apple II code from 40 years ago as an example. Russinovich wrote: "We are entering an era of automated, AI-accelerated vulnerability discovery that will be leveraged by both defenders and attackers."
In May 1986, Russinovich wrote a utility called Enhancer for the Apple II personal computer. The utility, written in 6502 machine language, added the ability to use a variable or BASIC expression for the destination of a GOTO, GOSUB, or RESTORE command, whereas without modification Applesoft BASIC would only accept a line number. Russinovich had Claude Opus 4.6, released early last month, look over the code. It decompiled the machine language and found several security issues, including a case of "silent incorrect behavior" where, if the destination line was not found, the program would set the pointer to the following line or past the end of the program, instead of reporting an error. The fix would be to check the carry flag, which is set if the line is not found, and branch to an error.
The existence of the vulnerability in Apple II type-in code has only amusement value, but the ability of AI to decompile embedded code and find vulnerabilities is a concern. "Billions of legacy microcontrollers exist globally, many likely running fragile or poorly audited firmware like this," said one comment to Russinovich's post.
Matt Schlicht et Ben Parr, les deux fondateurs du réseau social dystopique Moltbook qui permet à des instances IA de discuter entre elles, rejoignent Meta. Le groupe de Mark Zuckerberg rachète le réseau social par la même occasion.
Samsung says it's thinking about bringing "vibe coding" to future Galaxy phones, allowing users to describe apps or interface changes in plain language and have AI generate the code. TechRadar interviewed Won-Joon Choi, Samsung's head of mobile experience, to learn more about the plans. Here's an excerpt from their report: As noted by Won-Joon Choi, the usefulness of vibe coding on smartphones is that it opens up the "possibility of customizing your smartphone experience in new ways, not just your apps but your UX." He added, "Right now we're limited to premade tools, but with vibe coding, users could adjust their favorite apps or make something customized to their needs. So vibe coding is very interesting, and something we're looking into." [...]
Samsung recently debuted the Galaxy S26 series of phones and made a point to not call them smartphones -- they're "AI phones" now. This certainly rang true with the majority of upgrades to the devices being AI software-focused, like the new Now Nudge and expanded Audio Eraser tools, with the biggest hardware bump for the base models coming via the 39% improved NPU processing (the processor in charge of on-device AI tasks). It also teased the debut of Perplexity on its phones, joining as an alternative to the Gemini assistant, and teased the possibility of other AI models getting the same treatment in the future.
Dacia, le champion européen de Renault Group monte en gamme et muscle son jeu. Tout le catalogue s’enrichit de motorisations en phase avec le marché et sa clientèle n’hésite plus désormais à réclamer des équipements toujours plus modernes. La Spring ne fait pas exception, et s’éloigne en 2026 du modèle tout juste acceptable lancé en 2021. Nous l’avons prise en main dans les environs de Nice.
Un style bien plus moderne depuis 2024
On se rappelle de notre essai de la première Spring, nous étions en 2021. La première électrique de la marque n’avait quasiment rien pour elle, à commencer par son dessin. Son usage était pour le moins limité, avec un confort et des équipements qui nous renvoyaient vingt ans en arrière. Depuis, Dacia a revu sa copie, et la nouvelle mouture passe presque inaperçue dans la circulation depuis son gros restylage datant de quelques mois. Pas de surprise ici, le visage n’a pas changé avec une signature lumineuse à LEDs à l’avant, et un design globalement beaucoup plus moderne.
Pour autant, elle a toujours cette position sur la route un peu maladroite, notamment à l’arrière avec une carrosserie semblant surélevée sur des pneus fins. Elle n’a quand même pas l’air d’une voiture sans permis bodybuildée. Qu’on ne s’y trompe pas, on ne la prendra pas pour ce qu’elle n’est pas. Elle a toutefois ce petit quelque chose de rafraîchissant dans un paysage où les voitures sont de plus en plus grosses. Oui, elle semble cantonnée exclusivement à la ville. On note quelques petits détails appartenant à un passé pas si lointain, comme la tige servant d’antenne radio. Avec ses imposants pare-chocs et ses élargisseurs d’ailes en plastique brut, elle paraît parée aux chocs typiques d’une vie urbaine.
Un grand écran connecté
On connaissait déjà ce nouvel habitacle. Il reprend l’ambiance résolument moderne du Duster, avec du style et des technologies, sans trop en faire. On a ici un peu plus que l’essentiel, et des voitures plus chères en donnent même parfois moins. Le design de la planche de bord se montre plutôt valorisant. Bien sûr, il faut composer avec des plastiques durs, qui à défaut d’apporter une touche de luxe, lui confèrent une certaine robustesse. Rien n’est compliqué pour les usagers. Les commandes du bloc de climatisation ? Simples comme bonjour.
Un écran tactile de 10,1 pouces assure toute la connectivité que l’on attend d’une citadine en 2026. Dans cette finition, on peut compter sur Apple CarPlay et Android Auto, histoire de ne pas être dépaysé. Pour recharger ses appareils, on peut se reposer sur des prises USB-C. On adore le support téléphone si nécessaire, et bien sûr les très astucieuses attaches « YouClip », qui permettent d’accrocher ici et là différents accessoires disponibles. Sincèrement, on ne se sent pas si mal dans cette petite voiture. À l’arrière, on peut imaginer non pas voyager, mais emmener deux collègues pour aller au restaurant à 5 minutes du bureau. On est une fois de plus très étonné du coffre de 308 litres (+ frunk) de capacité dans cette Spring qui prend pourtant si peu de place.
100 ch qui changent tout !
On revient de loin ! La première génération de Spring n’avait que 45 chevaux sous le capot, et un couple de vélo électrique. On exagère à peine… Mais Dacia a mis maintenant le turbo, si l’on peut dire, avec désormais 100 chevaux pour le modèle « haut de gamme ». Clairement, ça change tout ! Niveau performances, on descend sous les 10 secondes pour atteindre les 100 km/h. Le constructeur d’origine roumaine aime communiquer sur la reprise 80 à 120 km/h, qui ne prend que 6,9 s. Et pour qu’elle puisse s’aventurer sur les voies express sans se traîner, la vitesse de pointe de 125 km/h suffit bien. On se surprend même à devoir regarder le compteur sur certaines portions de départementales, pour s’assurer de préserver notre permis.
Question autonomie, la fiche technique parle de 225 km. Pour cela, Dacia se repose sur une nouvelle batterie LFP (Lithium Fer Phosphate) de 26,8 kWh au lieu d’une NMC (Nickel Manganese Cobalt) pour une meilleure durée de vie et plus de sécurité. Les plus observateurs l’auront remarqué, ça n’ajoute pas plus de bornes pour autant, mais ça coûte moins cher. Car, on y reviendra, la voiture s’avère moins bon marché qu’auparavant. Sincèrement, on conduit désormais une voiture ayant une réactivité normale, prenant de la vitesse comme la plupart des citadines du moment, dont certaines bien plus grosses qu’elle. Entendons-nous bien, on n’évoque pas là des performances de GTi, mais d’une petite voiture à vocation urbaine, capable sur le papier de s’éloigner des villes.
Bien plus agréable à conduire
Justement, avant de partir, un regard sur les gommes nous surprend de la mauvaise façon. Les pneumatiques Linglong d’origine chinoise sont reconduits. Sauf que Dacia a travaillé sur son châssis, et la voiture n’a plus rien de la patineuse artistique, notamment sur le mouillé. Cette Spring de 100 ch sauce 2026 a droit à ce qui semble être la norme ailleurs, une barre antiroulis avant. S’il n’y avait que ça… On trouve aussi de nouvelles suspensions et plein d’autres petits ajustements. On ne conduit tout simplement pas la même voiture que nous avons connue à ses débuts. Elle ne s’avachit pas exagérément sur ses appuis, et garde raisonnablement bien son cap pour que l’on puisse parler d’une conduite plutôt sûre.
Si elle gagne en stabilité, elle a pris aussi un certain embonpoint, et dépasse désormais la tonne. À date, elle ne paie ainsi toujours pas le stationnement dans la capitale. N’oublions pas non plus qu’au passage, pour être commercialisée chez nous, elle s’équipe d’ADAS sophistiquées, celles tombant sous l’obligation de la norme GSRII. Maintenant, on ne risque pas non plus l’endormissement au volant de cette voiture qui, dans notre réalité n’atteint pas les 200 km, et moins encore si vous prenez l’autoroute. Si c’est nécessaire, la recharge en DC (40 kW) est accessible pour une recharge à 80 % en 30 minutes. Comptez sur moins de 3h30 sur une Wallbox de 7 kW. Petit bonus, le V2L que certains SUV bien plus onéreux ne supportent même pas.
Sans bonus, sa carrière se complique en France
Alors voilà, Dacia a modernisé sa Spring, et c’était plus que nécessaire pour espérer continuer à avoir une carrière commerciale honorable en Europe. Seulement, depuis on a sur le marché une Citroën ë-C3 bien moins chère, et une future nouveauté qui pourrait lui faire beaucoup de mal, la Twingo, elle aussi annoncée à un tarif bien plus compétitif. Le hic de la fabrication en Chine de la Spring pèse plus que jamais sur son prix, lui interdisant tout bonus, à l’exception de la maigre prime CEE. La gamme démarre sous la barre des 17 000 € et atteint les 19 700 € dans notre version huppée.
An anonymous reader quotes a report from the Guardian: AI has made it vastly easier for malicious hackers to identify anonymous social media accounts, a new study has warned. In most test scenarios, large language models (LLMs) -- the technology behind platforms such as ChatGPT -- successfully matched anonymous online users with their actual identities on other platforms, based on the information they posted. The AI researchers Simon Lermen and Daniel Paleka said LLMs make it cost effective to perform sophisticated privacy attacks, forcing a "fundamental reassessment of what can be considered private online".
In their experiment, the researchers fed anonymous accounts into an AI, and got it to scrape all the information it could. They gave a hypothetical example of a user talking about struggling at school, and walking their dog Biscuit through a "Dolores park." In that hypothetical case, the AI then searched elsewhere for those details and matched @anon_user42 to the known identity with a high degree of confidence. While this example was fictional, the paper's authors highlighted scenarios in which governments use AI to surveil dissidents and activists posting anonymously, or hackers are able to launch "highly personalized" scams.
Connexion
