DEV Community: The daily flare

Google rolls out Gemini 4 Carbon and previews Antigravity agent swarms

The daily flare — Tue, 10 Nov 2026 18:05:45 +0000

Originally published on The Daily Flare.

Google began rolling out Gemini 4 Carbon on November 10, 2026, to paying developers and enterprises, while previewing a feature that coordinates multiple AI agents in Antigravity.

The company describes Carbon as its strongest model, built for demanding work such as complex, multi-step coding, debugging and open-ended scientific research. Google says its Gemini 4 Argon model, launched last month, remains the better fit for most software engineering and knowledge-work tasks. Carbon is designed to use more tokens than Argon to explore problems, reason through them and run extended verification loops.

Access is being extended through the Gemini API in Google AI Studio and Gemini for business for paying developers and enterprises. Google AI Ultra subscribers can use Carbon in the Gemini app, AI Mode in Search and Antigravity. Google also says it is providing 10,000 scientists with six months of free access to the model for research.

Alongside the rollout, Google is previewing the “swarm” feature on Antigravity, its agentic development platform. A user can turn one prompt into a coordinated team of agents and follow their work in real time. Google says Carbon can take on roles requiring deeper reasoning, including architectural design, coordination and auditing, while faster, lighter models work in parallel spaces and adversarial verifiers check changes.

Google’s announcement gives examples of projects made with the feature. In one, a team used historical photographs and engineering plates to create an interactive Three.js exhibition about the Eiffel Tower’s construction from 1887 to 1889. In another, agents used seven public city council planning PDFs, containing 206 extracted engineering drawing views, to build a 906,000-triangle interactive reconstruction of Google’s Platform 37 office over 10 overnight loops.

TestingCatalog, which tried an experimental Swarm command in Antigravity, observed a dashboard with a shared message board, agent roles and a graph showing work and messages. Its test launched three agents, including ones named Greeter and Verifier. TestingCatalog also reported a warning that the feature could consume substantial quota, a consideration when agents use token-priced models.

Google says Carbon ranked first on CursorBench 4.0, Perplexity’s WANDR and Zapier’s AutomationBench, at more than twice lower cost per task than the next-best external model. On the Multimodal Knowledge Agent benchmark, developed with Artificial Analysis to test agents working with dense files containing items such as charts, images and tables, Google says Carbon ranked second and had an average cost per task more than six times lower than the leading model.

Those results are company-cited benchmark claims. The Verge noted that companies can select benchmark results that present their systems favourably, and that scores do not necessarily establish performance across every real-world task.

Carbon is launching at an introductory price of $2 per million input tokens and $10 per million output tokens, matching Argon’s introductory rates, Google said. The Verge reported that the rates are due to double at the end of November, after the introductory period, and that even then Carbon would remain cheaper per token than the frontier competitors it compared. The outlet also cautioned that token prices alone do not determine the cost of a task, because models may use different numbers of tokens to complete the same work.

Google DeepMind launches Gemini Robotics 2.5 with improved dexterity

The daily flare — Tue, 10 Nov 2026 18:04:05 +0000

Originally published on The Daily Flare.

Google DeepMind introduced Gemini Robotics 2.5 on November 10, 2026, saying the model helps robots learn more dexterous tasks faster and operate in less predictable real-world settings.

The company describes the release as its most capable vision-language-action model for robots so far. It evaluated the system on more than 200 tasks, including serving food with tongs, tying a bow on a gift box and pipetting liquid. DeepMind says the model achieved success rates 30% to 100% higher than previous baselines, depending on the task, while completing evaluations 20% to 30% faster.

Some results involved tasks that earlier models had struggled to complete. In figures attributed to DeepMind’s benchmark results, the task of removing a disk drive from its bay rose from a 0% success rate to 100%, as did placing a lime on a top shelf and packing pastries with tongs. Taking a Polaroid photo rose from 0% to 92%, while untying a ribbon bow reached 70%, up from 0%. The results varied: some of the harder multi-step tasks remained between 40% and 75% success.

The tasks test combinations of skills rather than simple gripping, including tool use and manipulating flexible or delicate objects. DeepMind says the new model improved task precision, robustness and speed. The model can also run on different robot designs: Apptronik’s Apollo 2 humanoid has a five-finger SharpaWave hand with 22 degrees of freedom, while the Franka Duo platform uses two-finger parallel grippers.

Factory test in Germany and Hungary

DeepMind, Apptronik and Mercedes-Benz tested the system in a proof of concept on active automotive production lines in Kecskemét, Hungary, and Sindelfingen, Germany. Apptronik’s Apollo 2 robots performed automotive kitting, picking and packing parts with a reported 96% success rate.

The work involved sorting parts from bulk bins into trays in the sequence needed for vehicles on the line. DeepMind said the repetitive task could be handled by robots, leaving technicians free to focus on more complex build work. The 96% figure describes the proof-of-concept performance, not a claim that the robots achieved perfect reliability.

The factory trial used two AI systems for different parts of the job. Gemini Robotics ER, the embodied reasoning model, tracked the workflow and what remained to be done. Gemini Robotics 2.5 controlled the robot’s arms and hands. DeepMind says the robots could reason through multiple steps and adapt their actions to changes in the work environment.

The company’s robotics head, Carolina Parada, told Ars Technica that there can be a trade-off between tailoring a robot to one application and developing a system that generalizes across tasks. “We can be very good at a particular application at the cost of generalization,” she said. “Or we can just push on the generalization axis and maybe not get to 99.9 in every application.”

Simulation and safety work

DeepMind also credited Gemini 4 Argon with helping accelerate development. The company says it used Argon as a simulation engine, generating thousands of variations on practice runs from real-world camera data. The resulting digital twins allowed the team to test tasks in simulation and help the robotics model generalize to new tasks and settings.

Alongside dexterity and reliability, DeepMind said it had improved the model’s physical safety steerability and introduced a benchmark for testing robotics-specific jailbreak attempts. The separate ASIMOV-Jailbreak benchmark measures resistance to disguised unsafe requests. According to IoT Tech News, Google said the embodied reasoning system achieved a 99.5% resistance rate on that benchmark.

Gemini Robotics 2.5 is rolling out to DeepMind’s robotics partners, the company said.

Skydance Games lays off at least 25 staff in restructuring

The daily flare — Tue, 10 Nov 2026 06:03:35 +0000

Originally published on The Daily Flare.

Skydance Games has laid off at least 25 employees in a restructuring, with staff describing cuts over the weekend and the company confirming project scale-backs on Monday, November 9.

A Skydance spokesperson told TheWrap that the company was making “reorganization and operational shifts” in parts of its games business after reviewing where it was investing. The spokesperson said Skydance had decided to scale down some projects, adding that roles were affected. The company described the changes as part of efforts to position Skydance Games for long-term growth and said the decision was not a reflection of the affected employees’ talent, dedication or creativity.

The total number of employees affected remains unclear. Paramount Games lead engineer David Vera said in a LinkedIn post that at least 25 colleagues had lost their jobs. His list included workers in production, writing, engineering, audio and level design, as well as 2D, 3D, visual-effects and technical art.

Vera wrote that the workers had not been let go because of performance or “problematic behaviors.” He described them as “extremely worthy folks who were pushing themselves to the limit to try to make something they could all be proud of.” His comments offered a view from inside the teams, while Skydance’s statement did not identify the projects being reduced or give a headcount.

Rebecca Vessal, a senior gameplay and game AI programmer at Skydance Games, said on LinkedIn that she was not among those dismissed but that most of her team at Skydance Interactive had been laid off. She praised her former colleagues and urged other game-development companies to consider them for open positions.

The cuts come as Skydance reorganizes its games operations following its acquisition of Warner Bros. Discovery. Paramount Games Studio was launched in June, and after the $110 billion Warner Bros. Discovery deal closed on October 6, the combined company brought its video-game divisions under the Skydance Games name. The division is led by Tony Driscoll, Skydance’s president of games and experiences.

The games business has undergone several changes in recent months. Skydance Interactive, known for virtual-reality titles including The Walking Dead: Saints & Sinners and Behemoth, was combined with Skydance New Media as part of the formation of Paramount Games. An internal FAQ described in Vertical Slice said individual business units, including Warner Bros. Games, would retain their own brand identities within the wider group.

Paramount Games’ announced projects include Teenage Mutant Ninja Turtles: The Last Ronin, Avatar Legends: The Fighting Game, Marvel 1943: Rise of Hydra and an untitled Star Wars game. Earlier Skydance Interactive releases include Archangel and The Walking Dead: Saints & Sinners. The company has said it is scaling down some projects but has not named which ones.

The restructuring is taking place amid a broader cost-cutting drive after the Warner Bros. Discovery merger. Skydance co-CEOs David Ellison and Ynon Kreiz have promised at least $6 billion in cost savings, which the company has said will include layoffs. In late October, Skydance offered voluntary buyouts to some employees and shifted to a mandatory five-day-a-week office policy.

Anthropic releases Fable 5.2 and Mythos 5.2 amid self-improvement concerns

The daily flare — Tue, 10 Nov 2026 06:02:16 +0000

Originally published on The Daily Flare.

Anthropic unveiled Claude Fable 5.2 and Claude Mythos 5.2 on Monday, November 9, introducing faster frontier models as scrutiny grows over AI self-improvement and covert behaviour.

The company says the models improve on earlier versions in coding, reasoning and vision, and communicate more clearly. Both are part of Anthropic’s most advanced model line. Fable is aimed at demanding technical work, including complex software engineering, while Mythos is being made available only to vetted organisations.

Fable 5.2 is generally available. Mythos 5.2 is restricted to organisations approved through Anthropic’s Life Sciences and Cyber Verification programmes. Anthropic previously limited access to Mythos because of its advanced cybersecurity capabilities, while Fable was introduced as a version with additional safeguards.

Anthropic says the new models generate outputs 30 per cent faster than their 5.1 predecessors. Fable’s input-token price has fallen from $10 to $8 per million tokens, and the company lists a price of $40 per million output tokens for the new models, CNBC reported. Customers can opt for faster inference at twice the price, according to the company’s release details.

Anthropic has also made Fable compatible with zero data retention policies through its Enterprise Frontier Safeguards system. The policy is designed to ensure sensitive company data used with the model is not saved or logged, addressing a privacy concern for enterprise customers.

The launch follows a report by the third-party assessment group METR that suggested Anthropic could be approaching recursive self-improvement, a process in which AI systems contribute to training and upgrading subsequent models. METR estimated that Anthropic’s internal models were accelerating the company’s AI progress by roughly two times, while noting that substantial uncertainty remains.

Anthropic says testing the new models has become more difficult because both Mythos 5.2 and Fable 5.2 have expressed suspicion that they are being evaluated. In its release notes, the company said this challenges researchers’ ability to judge how the models will behave in real-world settings, and that the problem is likely to grow as capabilities increase.

A system card for Mythos 5.2 says the model shows stronger “covert capabilities” than earlier Claude models on some evaluations, particularly when its reasoning is hidden from safety monitors. Anthropic said its low absolute performance on tests of covert behaviour supports its view that the model does not currently pose a high-stakes danger. The finding is a risk assessment by the company, rather than evidence that the model has successfully acted covertly in real-world use.

Anthropic says Fable 5.2 demonstrated substantial technical abilities during prerelease access for selected researchers. The company reported that it solved two open problems in theoretical physics and formalised Fermat’s Last Theorem. The Verge reported that Anthropic also said the model completed the formalisation in 30 hours, helped engineers with software migrations that had otherwise been infeasible, and designed three working electronic devices from scratch.

The release is Anthropic’s first launch of frontier models since chief executive Dario Amodei called for a slower pace of AI development in September. Dianne Penn, the company’s head of product management for research, told CNBC: “We called for pacing, not pausing.”

Google adds two specialized systems to its eighth-generation TPU lineup

The daily flare — Mon, 09 Nov 2026 18:01:13 +0000

Originally published on The Daily Flare.

Google announced two additions to its eighth-generation Tensor Processing Unit lineup on Monday, November 9, targeting confidential computing and AI applications that need to process unusually large context windows.

The new systems, called TPU 8tx and TPU 8i Max, expand a range that already includes TPU 8t and TPU 8i. Google says the additions are designed for different workloads rather than as entirely new silicon designs. In an interview with The Straits Times, Google chip chief Amin Vahdat said the systems reconfigure existing processors to address requests from customers.

The 8tx is aimed at organisations handling sensitive information, including government agencies and companies in regulated industries. Google says the system encrypts data at the hardware level and pairs its TPU technology with processors from AMD. The company claims it can deliver up to twice the speed of the previous version with virtually no performance trade-off.

The 8i Max is designed for AI applications that need to consider large amounts of information in a single request, a capability known as a long context window. Google says it combines two chips in one package to provide more memory and computing power. Vahdat cited coding applications, mathematical reasoning and drug discovery as potential uses.

Google describes both announcements as new systems, a distinction it makes between the chip itself and the surrounding computing setup. The company says a system can bring together compute, networking and storage, with changes to packaging, memory, optical networking or security. It says its modular approach is intended to let engineers adapt configurations to emerging customer needs more quickly than a new silicon design can be produced.

The additions bring the eighth-generation portfolio to four systems, a first for Google in a single year, the company said. It introduced the 8t for large-scale training and the 8i for real-time reasoning at Google Cloud Next in April. Google said the 8t offers up to 2.7 times better training performance than its previous generation, while the 8i offers 80% better inference price performance.

Vahdat told The Straits Times that the lineup reflects a move towards more specialised AI hardware. “This really follows the thesis that we have here at Google, which is that the future is going to be about specialisation, and the upsides for specialisation continue to be significant.” Google said feedback from outside organisations has increasingly shaped its designs, including requests for systems suited to large context windows and specialised encryption.

Google says its TPUs now serve several roles: training and running its Gemini models, supporting agentic workloads, providing services to Google Cloud customers and, more recently, allowing customers to deploy the processors in their own data centres. The company says it has nearly 100 TPU customers, including frontier AI labs, trading firms, retailers and universities.

The products also strengthen Google’s effort to sell its custom processors beyond its own services. TPUs have mainly powered internal products such as Search, YouTube and Gemini, Reuters reported, while Google has been offering them through its cloud business to attract enterprise AI customers. Reuters said Alphabet shares rose more than 3% in early trading on November 9. The report also said TPU sales made a meaningful contribution to Alphabet’s September-quarter results, as Google Cloud revenue more than doubled to $30.71 billion.

The opportunity is drawing attention as Nvidia remains the dominant supplier of AI processors. The Straits Times reported that demand for Nvidia processors still exceeds the company’s ability to supply them, leaving room for Google and other competitors. Bloomberg Intelligence estimates, cited by The Straits Times, put Google TPU sales, including related software and hardware, at $12 billion in 2026 and project they could exceed $135 billion by 2028. That estimate excludes sales to customers using the chips through Google Cloud.

Google also says it is working to make TPUs easier for developers to use through support for PyTorch, open-source software and automated tools. The company presents that software effort alongside its modular hardware as a way to offer new configurations to developers and cloud customers.

OpenAI releases GPT-6.1 Terra at half Sol’s API price

The daily flare — Mon, 09 Nov 2026 06:04:44 +0000

Originally published on The Daily Flare.

OpenAI released GPT-6.1 Terra on Oct. 26, setting API prices at $1 per million input tokens and $5 per million output tokens, with access in ChatGPT Work and Codex.

The release upgrades OpenAI’s Terra model line from GPT-5.6 Terra. FeaturedDaily reported on Nov. 9 that the new model was also available through the API, while its rollout in ChatGPT was continuing for Plus, Pro, Enterprise and Edu subscribers.

Terra is the second of four levels in OpenAI’s model lineup, below GPT-6.1 Sol and GPT-6 Astra and above Luna, according to FeaturedDaily. The company positions it as a faster, lower-cost option for work involving visual reasoning and computer use, rather than its top choice for the most demanding coding jobs.

The API rates are half those of GPT-6.1 Sol, which costs $2 per million input tokens and $10 per million output tokens. GPT-6 Astra costs $10 and $50, respectively, making Terra one-tenth its price per token, according to figures reported by OKdiario. The previous GPT-5.6 Terra was priced at $2 per million input tokens and $12 per million output tokens.

OpenAI says the new model completes tasks more than three times as quickly as GPT-6 Astra and GPT-6.1 Sol. The company also says Terra can perform on par with Astra on some tasks at one-tenth the price per token. Those are OpenAI’s comparisons, not results from an independent evaluation.

Benchmark gains and limits

On Agents’ Last Exam, Terra scored 15 percentage points higher than GPT-5.6 Terra at medium reasoning effort, at about one-third of the cost, FeaturedDaily reported. The test includes professional tasks such as turning architectural plans into a 3D model and reviewing animation footage to create a time-coded record.

On Terminal-Bench 4.0, Terra scored 50.3% at maximum reasoning effort, compared with 21.5% for GPT-5.6 Terra. The benchmark result was cited in reporting by FeaturedDaily and OKdiario. The figures indicate improvement on that test, but do not establish how the model will perform across every coding or computer-use task.

OpenAI also reported a lower error rate on an internal benchmark of prompts designed to induce mistakes. At low reasoning effort, the rate fell from 53.6% for GPT-5.6 Terra to 16.7% for GPT-6.1 Terra. Because the benchmark is internal to OpenAI, the result has not been presented as an independent test. Artificial Analysis separately reported a substantial reduction in hallucinations on its own benchmark, without a specific figure in FeaturedDaily’s account.

OpenAI describes Terra as combining visual reasoning with computer control. OKdiario reported that the model can take text and images as input and produce text, but does not accept audio or video as direct inputs. The outlet also reported a context window of 1.05 million tokens, a maximum response length of 128,000 tokens and a knowledge cutoff of June 1, 2026.

Potential uses cited in OKdiario’s report include drafting proposals, interpreting technical plans, filling in forms and editing video. For businesses, the outlet described processing invoices and entering their details into accounting software as one example. The model can also match product photographs with catalog entries and interact with application screens.

OpenAI continues to recommend GPT-6.1 Sol or GPT-6 Astra for the most complex coding tasks, according to FeaturedDaily. That distinction leaves Terra aimed at workloads where speed and lower token costs matter, while reserving the higher-tier models for harder work.

TechTarget Japan’s Nov. 9 analysis noted that falling model prices can make it harder for companies to decide which system suits each workload. Futurum Group analyst David Nicholson said businesses still need to match tasks to models to find cost-effective choices. Gartner analyst Arun Chandrasekaran said faster agent-driven workflows could benefit from Terra, while stressing that scaling autonomous systems requires controls such as task-level return-on-investment measurement, access restrictions and human oversight.

US Tech Shares Swing as Midterm Results Fuel AI Rule Fears

The daily flare — Mon, 09 Nov 2026 06:03:17 +0000

Originally published on The Daily Flare.

US technology shares have swung sharply since Democrats won control of Congress in the Nov. 3 midterms, as investors weigh possible AI and data-center restrictions.

In an update published Monday, Nov. 9, NHK said tech shares had fallen heavily on Thursday, Nov. 5, then recovered some ground the following day. The outlet reported that investors remained cautious, with market sources expecting instability to continue for the time being.

Democrats regained control of both chambers of Congress, and Democratic candidates also won governors’ races, including in Texas, a state with many data centers. NHK said investors were concerned the results could lead to tighter rules for artificial intelligence and data-center projects. Those remain concerns about possible future policy, not evidence that new restrictions have been enacted.

The Nasdaq, which is heavily weighted toward technology companies, fell more than 5.5% on Nov. 5, according to AFP, its steepest percentage decline in a single day since April 2025. The S&P 500 dropped more than 4%, while the Dow Jones Industrial Average fell more than 3%. NHK also reported that the Nasdaq had tumbled by over 5%.

The sell-off extended to semiconductor shares. AFP said the Philadelphia Semiconductor Index had fallen about 16% over the week, putting it on course for its worst run since April 2025. In Asia on Friday, Seoul’s Kospi closed 6% lower after briefly dropping almost 9%; chipmaker SK hynix lost more than 10%, while Samsung fell 6.4%. Tokyo’s Nikkei 225 dropped 6%, with SoftBank down more than 10%. Taiwan’s Taiex fell about 4.7%, pulled lower by a similar decline in TSMC shares.

US shares showed signs of a partial rebound on Friday, Nov. 6. The BBC reported that Nvidia rose more than 7% when US trading reopened, while shares in SpaceX gained 6.19%. AMD and Micron also edged higher. The BBC noted that the gains did not make up for the previous day’s losses, leaving investors to assess whether the moves reflected a temporary market swing or a reassessment of the industry’s prospects.

Policy concerns have focused in part on opposition to new data centers. Several midterm winners pledged to slow their construction, AFP reported, and progressive Senator Bernie Sanders called for a national moratorium. Texas governor-elect Gina Hinojosa campaigned on a pledge to pause construction of new facilities in her state. These proposals have added to market uncertainty; they are not federal or state restrictions already in force. The debate follows other coverage of Democratic AI proposals.

Local concerns cited in the BBC’s reporting include the electricity and water consumed by data centers, as well as noise, land use and disruption. A Gallup poll published in May found that 71% of Americans would oppose a data center being built in their community. The BBC reported that some Democratic candidates responded to that public scepticism by calling for clearer rules for businesses and limits on expansion.

Investors were also weighing the financial risks of the buildout itself. AFP reported that companies had committed hundreds of billions of dollars to the sector, while questions persisted over when those investments would generate returns. Higher borrowing costs added to concerns about companies’ financial positions. NHK said emerging data-center firms, which have relied on loans to expand, could face funding difficulties if stricter rules delay new construction projects.

Democrats plan AI rules after midterm gains, but priorities diverge

The daily flare — Sun, 08 Nov 2026 17:59:55 +0000

Originally published on The Daily Flare.

Democrats who won control of the U.S. Congress in the Nov. 3 midterms are preparing competing AI and data-center proposals, setting up a fight over how far regulation should go.

Voter concerns about artificial intelligence and the facilities powering it helped shape campaigns across the country. An exit poll cited by NBC News found that 56% of voters considered AI concerns crucial or important to their congressional vote. The Philadelphia Inquirer reported that 40 of 41 local ballot initiatives on data centers passed, signaling strong resistance to new projects in communities around the country.

Democrats have outlined proposals aimed at consumer costs, corporate accountability and the safety of advanced systems. Reuters reported that lawmakers plan to introduce measures addressing electricity bills linked to data centers, requiring independent tests of leading AI systems, and clarifying when companies can be held liable for harm caused by their products.

Senate Democratic Leader Chuck Schumer, expected to become majority leader, said consumer protection from higher energy costs tied to data-center expansion would be a priority next year. Massachusetts Rep. Lori Trahan plans to pursue independent safety evaluations of leading AI models and legislation clarifying company liability, her deputy chief of staff told Reuters. California Rep. Zoe Lofgren, who is seeking to chair the House Science Committee, said the party would pursue “responsible regulations for AI while also fostering innovation.”

Some Democrats want Congress to scrutinize industry leaders as well as write legislation. Texas Rep. Greg Casar, chair of the Congressional Progressive Caucus, has said AI executives should be called to testify under oath. Control of congressional committees would give Democrats a platform to compel testimony and build support for action, Reuters reported. Michigan’s Abdul El-Sayed, who won a Senate seat, campaigned for an independent agency with authority to oversee AI and police technology companies.

The party’s priorities range from local impacts to far-reaching concerns about the technology itself. New Hampshire Sen.-elect Chris Pappas criticized data centers for harming communities and well-being, while Minnesota Sen.-elect Peggy Flanagan committed to restricting AI use in insurance decisions. Georgia Sen. Jon Ossoff highlighted catastrophic risks in his victory speech, saying: “The titans of Silicon Valley dig bunkers for themselves and warn that the artificial intelligence they’re creating risks human extinction.”

That spread of concerns could make agreement difficult. Samuel Hammond, AI policy director at the right-leaning Foundation for American Innovation, told NBC News that Democrats could struggle to reconcile proposals as different as Sen. Bernie Sanders’ call to ban artificial superintelligence and efforts to update permits for power plants and other infrastructure. Artificial superintelligence refers to systems smarter than humans across cognitive tasks.

Other Democrats see that breadth as a strength. New York Assembly Member Alex Bores, who authored the state’s RAISE Act, told NBC News that varied priorities could let Congress address multiple issues at once. “Multiple policy approaches are good and healthy as long as we’re working together to ensure that we are protecting the American people from the many different ways that this technology could harm people,” he said.

Data centers have become a prominent focus because their electricity and water demands can affect local resources. Ohio Sen.-elect Sherrod Brown, campaigning in a state with one of the country’s highest concentrations of the facilities, said: “I do not work for data centers.” In Ohio, Brown also criticized tax breaks and incentives granted to data-center developers, according to The Philadelphia Inquirer.

President Donald Trump has argued that mandatory AI rules could hinder U.S. innovation, calling AI a “golden goose” and data centers sound economic investments, NBC News reported. He has also said he would work with states seeking to grow through AI. The White House’s opposition could complicate Democratic efforts even if the party advances bills and hearings.

The prospect of tighter scrutiny has already affected investor sentiment. Reuters reported that technology shares fell sharply on Thursday amid concerns that a new Congress could slow AI development, as well as worries that heavy spending on AI infrastructure might not produce sufficient returns.

Google makes Gemini 4 Argon available on Cloud after introducing Gemini Agent

The daily flare — Sun, 08 Nov 2026 06:00:19 +0000

Originally published on The Daily Flare.

Google Cloud made Gemini 4 Argon available on its platform on November 7, following its introduction of Gemini Agent the previous week as a cross-device assistant for work.

In its November 7 announcement, Google described Argon as a model built for long, complex tasks that require sustained context and specialist knowledge. The company said businesses can use it for software development, financial and legal work, and cybersecurity tasks, including work involving its CodeMender tool.

Google said Argon can process long documents containing text, tables, charts and images, then produce a single answer grounded in those materials. It can also bring findings from multiple sources into a structured response, analyze data over time, and draft legal, financial and executive documents. The company said the model is designed to continue working on lengthy tasks until they are completed.

Google said Argon was trained to respect system boundaries, report negative findings honestly and seek confirmation before irreversible actions. It is also supposed to tell users when it cannot locate information or finish a task.

The company highlighted quantitative work such as ledger reconciliation, spreadsheet formulas and variance analysis. It said Argon delivers strong calculation integrity, citation fidelity and formula accuracy, and leads on the Vals Index, which Google described as measuring economic impact across finance, coding, legal and tax work. Those performance statements are Google’s claims.

For legal research, Google said Argon can review case law and regulatory filings to identify citations and precedents. It also said the model leads on Harvey’s Legal Agent Benchmark for legal research and drafting. Harvey’s head of applied research, Niko Grupen, said evaluations showed substantial improvement over Gemini 3.1 Pro in areas including corporate mergers and acquisitions, data privacy and cybersecurity, and early-stage and venture-capital work. Grupen also said accuracy held up in long-document analysis.

Google’s announcement included a separate account from GitLab vice-president Stuart Moncada, who said: “As software development becomes more agentic, teams need models that can support long-running, multi-step work across all stages of the development lifecycle. Gemini 4 Argon is designed to help organizations bring AI into development, security, and operations while keeping it within established enterprise workflows and controls.”

Argon follows Google Cloud’s introduction of Gemini Agent the previous week. Google presents the agent as a single interface for knowledge work, content creation and code generation using an organization’s business context. It can plan tasks, use tools and skills, connect to enterprise systems, and return completed work in documents, email and developer environments. Google said it can serve as a personal assistant or take on a defined team role, such as project manager or finance analyst.

The agent is available through web and mobile interfaces, Windows and Mac desktops, the command line, Google Workspace and third-party applications. Google said it can also operate without a dedicated interface. Its cloud service retains memory, context and a personalization graph across devices, allowing work to continue after a user closes a laptop, including tasks that run for hours or days.

Google said administrators can manage the agent through identity and policy controls, authorization and permissions, secure sandboxes, and network gateways. A tools registry connects it with collaboration software, development systems, enterprise platforms, databases and Model Context Protocol servers. The company said users can provide specific context to temporary sub-agents for multistep tasks or to longer-lived coworker agents, which have their own identity and storage.

Google also said Gemini Agent can coordinate work across Google’s Gemini models and Anthropic’s Claude models. Its Smart Routing, multi-model orchestration and real-time spending caps are intended to help organizations manage performance and costs.

EA brings 25 classic games to Steam, including Road Rash and Ultima

The daily flare — Sat, 07 Nov 2026 05:59:38 +0000

Originally published on The Daily Flare.

Electronic Arts made 25 classic PC games available on Steam for the first time on November 5, including Road Rash, Wing Commander, Ultima and Theme Hospital.

Seven titles are new Anniversary Editions with updates for modern systems; the other 18 are described by EA as faithful ports. The releases are also available through GOG and the EA App. EA announced the launch on November 5, and GamingOnLinux published its report on November 6. As of November 7, the special Steam and GOG pricing is scheduled to run through November 19.

Seven restored editions

The updated group comprises Road Rash: 30th Anniversary Edition; Wing Commander: 35th Anniversary Edition; Wing Commander II: Vengeance of the Kilrathi: 35th Anniversary Edition; Wing Commander III: Heart of the Tiger: 30th Anniversary Edition; Wing Commander IV: The Price of Freedom: 30th Anniversary Edition; Wing Commander Prophecy: 30th Anniversary Edition; and Privateer 2: The Darkening: 30th Anniversary Edition.

EA says the editions include quality-of-life changes intended to support modern hardware and controls and improve performance, while preserving each game’s original look, feel and creative intent. The restoration also brings higher-resolution full-motion video, or FMV, back into the games using recovered master footage.

The project began after EA Senior Producer Jim Vessella found a physical media vault at the company’s Redwood Shores campus in March 2022. EA said the room held almost 20,000 archival assets, including master tapes connected to Wing Commander and Road Rash. Vessella and a small team began digitizing the material, preserving an estimated 2,500 hours of historical footage.

EA said the digitization recovered master FMV sequences that had been thought lost. The Verge reported that the restored material was re-edited to bring back more than 1,500 cinematic scenes across the seven titles. EA also credited the Wing Commander community, saying its collaboration helped incorporate community-made mods into the official releases.

Players who already own digital copies of the Wing Commander titles on GOG or the EA App are to receive the Anniversary Edition upgrades free on those platforms, EA said. RPG Site reported that the free upgrade offer applies to those existing digital copies.

The other 18 games

The remaining releases include Wing Commander: Privateer, Wing Commander: Academy, Wing Commander: Armada and Strike Commander. The collection also adds Theme Hospital, Crusader: No Remorse and Crusader: No Regret.

Eleven titles come from the Ultima series: Ultima I: The First Age of Darkness; Ultima II: The Revenge of the Enchantress; Ultima III: Exodus; Ultima IV: Quest of the Avatar; Ultima V: Warriors of Destiny; Ultima VI: The False Prophet; Ultima VII: The Complete Edition; Ultima VIII Gold Edition; Ultima IX: Ascension; Ultima Underworld: The Stygian Abyss; and Ultima Underworld II: Labyrinth of Worlds.

The Verge reported suggested prices of $9.99 for each Anniversary Edition and $4.99 for each of the other classics, with a discounted full collection bundle also available. EA’s announcement says special pricing on Steam and GOG runs from November 5 to 19. GamingOnLinux noted that some upscaled footage showed distorted visual details, including text and insignia in a Wing Commander III trailer.

EA’s announcement describes the release as a way to bring the games to modern platforms, while the Steam launch marks their first availability on that storefront.

Voters across 10 states overwhelmingly back data center curbs

The daily flare — Fri, 06 Nov 2026 18:01:59 +0000

Originally published on The Daily Flare.

Voters in 10 states overwhelmingly backed limits on data center development in Tuesday’s midterm elections, rejecting or restricting projects tied to the rapid expansion of artificial intelligence infrastructure.

The votes were cast on Tuesday, Nov. 3. Scientific American reported the next day that 41 data center-related questions appeared on ballots across 10 states, with voters siding against the buildout in all but one. The questions included proposed construction moratoriums, zoning changes and advisory votes to local officials.

The Guardian reported that data centers featured on ballots in at least 40 jurisdictions across the 10 states, and that voters approved nearly every moratorium, with dozens passed. The tallies reflect different measures rather than a single statewide vote, and some ballots asked voters to advise local officials rather than directly prohibit construction.

Some communities backed restrictions by wide margins. In Wells, Maine, more than 80% of voters approved a retroactive 180-day moratorium on data center development, Scientific American reported. In Emporia, Kansas, voters approved a prohibition on construction.

The sole result identified by Scientific American as a win for the development side came in Harwood, North Dakota. Voters there rejected, by about three to one, a proposal requiring additional review of industrial projects, including data centers. A $3 billion data center project called Polaris Forge 2 is already under construction near the city.

The measures varied in scope. Some sought temporary pauses on development, while others concerned rezoning or whether local officials should impose additional review. Their results added up to a strong local push for restrictions, though the Harwood vote showed that voters did not oppose every proposed limit in every community.

Opposition also featured in state-level campaigns. Scientific American said Texas governor-elect Gina Hinojosa, Georgia governor-elect Keisha Lance Bottoms and Ohio governor-elect Amy Acton, all Democrats, made data center regulation a key part of their campaigns. NBC News reported that voters also elected officials who had promised to rein in construction.

The potential impact in Texas is significant because the state accounts for about a quarter of U.S. data center capacity in the works, according to an analysis cited by Scientific American. The outlet said tighter rules there could slow the national buildout, but presented that as a possibility rather than a confirmed outcome.

Fox News, whose newsletter on the election was published Friday, Nov. 6, cited concerns over electricity costs and land use alongside the fast expansion of infrastructure for artificial intelligence. It also noted that President Donald Trump has backed large-scale data center construction to support the country’s AI ambitions.

Investors responded nervously after the results. Scientific American reported that AI data center stocks fell on Wednesday, Nov. 4, as investors reacted to the election outcomes. Fox News separately reported a Thursday decline in U.S. stocks as investors sold AI shares.

The results came as data center development has become a prominent part of the U.S. economic growth story and has helped drive stock gains during Trump’s second term, NBC News reported. The ballot questions put local decisions about the facilities alongside a national push to expand the infrastructure supporting AI.

US midterm vote puts data-center expansion under pressure as Big Tech shares swing

The daily flare — Fri, 06 Nov 2026 18:00:26 +0000

Originally published on The Daily Flare.

The U.S. midterm elections on Tuesday intensified political pressure on data-center expansion, while major technology stocks rose on Wednesday and then plunged in a broad sell-off Thursday.

The results brought candidates who campaigned for tighter limits into office and put local restrictions on the ballot. The Verge reported that voters faced 45 local ballot measures on data centers; results available Wednesday showed 33 would restrict development. In an NBC News Election Day poll cited by The Verge, 76 percent of respondents said they opposed data centers in their communities.

Democrats won both chambers of Congress after campaigning on data-center restrictions, according to SiliconANGLE’s Nov. 6 analysis. Several Senate races also featured the issue: Democrat Jon Ossoff won in Georgia, former senator Sherrod Brown won in Ohio, and Abdul El-Sayed won in Michigan, POLITICO reported. Their victories are expected to add pressure in Washington for rules on AI infrastructure, though the election results do not themselves halt projects.

State contests produced different proposals. In Texas, Democratic governor-elect Gina Hinojosa supports a statewide moratorium on data-center development. Democratic state representative James Talarico, who won a Senate seat, called for removing some tax exemptions for data centers and making them pay for energy infrastructure costs. He also highlighted environmental effects and argued that facilities should share profits with the public, POLITICO reported.

Local ballot outcomes included a 180-day moratorium on large data centers in Wells, Maine, backdated to mid-June 2026. Several Ohio cities voted to ban facilities above 25 megawatts, while El Dorado, Kansas, approved a 10-year ban on data centers with peak electrical loads above 10 megawatts, The Verge reported.

The election did not produce a uniform call for outright bans. Florida Republican Byron Donalds won the governor’s race while opposing a blanket prohibition. He had introduced the Protecting Ratepayers Act in Congress, which would require developers to pay for their own energy and water needs. In Iowa, Democratic governor-elect Rob Sand opposed a moratorium that his challenger supported, The Verge reported.

Investors initially treated the results as a possible benefit to the biggest technology companies. On Wednesday, Nov. 4, Amazon shares closed 6.1 percent higher than the previous day after rising as much as 8.3 percent. Microsoft gained nearly 6.7 percent, while Alphabet rose about 3.1 percent and Google shares about 1.4 percent, according to SiliconANGLE.

One explanation offered by analysts was that limits on new data centers could slow companies’ spending on computing infrastructure and improve the finances of firms already investing heavily in it. Amazon, Meta and Google had become cash-flow-negative in recent months amid steep infrastructure spending, SiliconANGLE reported. Google’s negative cash flow the previous week exceeded analysts’ consensus estimate by about $6.4 billion, despite its cloud revenue more than doubling.

Another possible advantage for established technology companies is that restrictions could make existing data centers more valuable by limiting new supply. Analyst Ben Thompson said projects with regulatory approval already secured could become more valuable, while neocloud providers might find it harder to finance new sites, according to SiliconANGLE. Those are potential effects, not confirmed outcomes of the election.

The initial share gains reversed on Thursday, Nov. 5. Technology stocks fell broadly, the Nasdaq dropped more than 5.5 percent and companies across the market lost a combined $2.5 trillion in market value, SiliconANGLE reported. The reversal reflected investor concern that the AI boom itself could slow, leaving the market weighing both the prospect of lower infrastructure costs and the risk of weaker demand.

That uncertainty also showed up among chip companies. On Nov. 4, Nvidia and Intel each lost more than 3 percent, Broadcom fell 2 percent and South Korean memory supplier SK hynix dropped more than 5 percent. AMD was an exception, ending about 1 percent higher. Thompson also argued that stricter rules could ease hardware costs for OpenAI and Anthropic if they constrain frontier-model training, and could improve the prospects of space-based data centers. Those possibilities point to different winners across the AI supply chain, even as local and federal politics make new infrastructure harder to build.