France's lower house votes 11-10 to tax internet traffic from Google, Netflix and Meta at about 6 US cents a gigabyte

The stories

14 stories on this beat today.

Every story told the same way: what happened, who is involved, why it matters, how it works, the truth and the good news.

1. 01

   France's lower house votes 11-10 to tax internet traffic from Google, Netflix and Meta at about 6 US cents a gigabyte

   €0.055

   tax per gigabyte sent to French users, about 6 US cents

   set in Laurent Lhardit's amendment to the budget bill [1]

   11 to 10

   the National Assembly vote that put it in the budget bill

   held on 30 October, against the French government's wishes [1]

   €10 billion

   a year, the amendment's own estimate of what it would raise

   each firm's bill is capped at 15% of its French sales [1]

   What happened

   • France's National Assembly voted 11 to 10 for a tax on internet traffic on 30 October. [1]
   • Firms that send data to French users would pay €0.055, about 6 US cents, per gigabyte. [1]
   • The vote adds the tax to next year's budget bill, but more votes are to come. [1]
   • French internet providers such as Orange would measure how much data each company sends. [1]
   • They would report the totals to France's tax office every three months. [1]
   • Arcep, France's telecoms regulator, would oversee the counting. [1]
   • Lhardit estimates the tax would raise about €10 billion a year. [1]
   • About 200 billion gigabytes reach French users through their providers each year. [1]
   • Five big companies named in the amendment send about half of that traffic. [1]
   • Each company's bill is capped at 15% of its French sales or 2% of worldwide sales. [1]
   • The finance committee's lead member argued that firms would pass the cost on to users. [1]
   • Lhardit replied that Facebook, Instagram and TikTok are free to use. [1]
   • "I fail to see how a tax can be passed on," he said of these apps. [1]
   • Opponents also warn that counting traffic this way could put users' privacy at risk. [1]

   Who is involved

   Laurent Lhardit, a Socialist member of France's parliament, wrote the change. [1]

   His notes name Google, Meta, Netflix, Amazon and Akamai as likely payers. [1]

   Akamai runs networks that keep copies of websites close to users, so pages load faster. [1]

   France's government opposed the change, and so did the finance committee's lead member. [1]

   Minister David Amiel doubted that traffic could be traced to each firm without complicated monitoring. [1]

   • Laurent Lhardit a Socialist member of France's National Assembly; he wrote the amendment
   • Google
   • Meta
   • Netflix
   • Amazon
   • France's government
   • Finance committee's lead member
   • Arcep France's telecoms regulator; it would oversee the counting
   • Facebook
   • Instagram
   • TikTok
   • David Amiel a French government minister; he doubted the traffic could be counted fairly
   • A few big senders about half the traffic
   • Google, Meta, Netflix, Amazon and Akamai big US internet firms named as likely payers; together with others they send most of France's traffic

   How it unfolded

   Several European countries already tax big tech firms on their advertising and marketplace sales. [1]

   In July Meta began charging advertisers extra fees in France, Italy and Spain. [1]

   Meta said the fees cover such taxes and other government charges. [1]

   1. March Meta announces fees
   2. July fees start in France, Italy, Spain
   3. 30 Oct Assembly vote, 11 to 10
   4. Next further votes
   5. July Meta's fees for advertisers start in France, Italy and Spain [1]
   6. 30 Oct the National Assembly votes 11 to 10 for the traffic tax [1]
   7. Next further votes before the budget becomes law [1]

   Why it matters

   If it passes, the biggest bills land on the few firms sending most of the data. [1]

   How it works

   Economists call it a per-unit tax: the bill grows with gigabytes sent, not with profit. [1]

   The truth

   • What is really going on France's lower house voted 11 to 10 to tax Google, Netflix, Meta and others for the data they send to French users, though France's government opposes it and more votes remain. [1]
   • Why it works on people A big round number, like €10 billion a year, makes a first vote by a single seat sound like a law already passed.
   • France's budget would gain about €10 billion a year, if the author's estimate holds. [1]
   • Nobody yet knows whether users would end up paying: the two sides disagree. [1]
   • Who gains France's state budget — Would collect about €10 billion a year by the amendment's own estimate, mostly from global technology companies. [1]
   • Who pays Google, Meta, Netflix, Amazon and Akamai — Would be billed €0.055 for each gigabyte sent to French users, up to 15% of their French sales. [1]
   • Nobody knows yet

     Whether France's traffic tax survives the rest of the budget process.

     The vote was 11 to 10, France's government opposes it, and further parliamentary votes are still to come. [1]

   • Nobody knows yet

     Whether the cost would reach users.

     The finance committee's lead member says firms would pass it on; its author says free apps cannot. [1]

   • Nobody knows yet

     How the traffic would be counted without watching users closely.

     Minister David Amiel doubted it can be traced to each firm without complicated monitoring. [1]

   What comes next

   The amendment must still survive further votes in France's parliament before it becomes law. [1]

1. 02

   OpenAI has $11bn of the $20bn it wants

   What happened

   • OpenAI, which makes ChatGPT, now aims to raise up to $20 billion in its new round. [3]
   • It first sought $30 billion; the Financial Times reported the $10 billion cut on Friday. [2]
   • So far it has secured about $11 billion, The Wall Street Journal reported. [4]
   • Interest from some large investors was tepid, the Journal said. [4]

   Who is involved

   Investors from the United Arab Emirates, including the state-backed fund MGX, should be the largest backers. [2]

   OpenAI will next ask family offices, which manage one rich family's money, and wealthy individuals. [2]

   • UAE investors including MGX, a state-backed fund
   • Family offices
   • Wealthy individuals
   • Anthropic OpenAI's main rival

   How it unfolded

   The round values OpenAI at $1.25 trillion, down from the $1.4 trillion it first pitched. [2]

   AI shares fell all week; Alphabet's dropped about 12% on Thursday, its worst day on record. [5]

   Why it matters

   OpenAI has told investors it plans about $600 billion of computing spending by 2030. [6]

   Anthropic, its main rival, plans to sell its shares to the public next month. [2]

   How it works

   Venture rounds can close in stages: money signed by a deadline, then more later. [2]

   The truth

   • What is really going on Elsewhere, OpenAI has secured only about $11 billion of the $20 billion it now wants, as investors ask whether AI spending will pay back. [2][3]
   • A later offer will let wealthy individuals buy at the same $1.25 trillion value. [2]
   • The round is still open, so the final size could change, CNBC's source said. [3]
   • Nobody knows yet

     How much OpenAI's round finally raises.

     The round is still open, the first close is expected below $20 billion, and OpenAI did not immediately respond to requests for comment. [3][4]

1. 03

   The Pope asks states to ban killer robots

   What happened

   • Pope Leo XIV called on Friday for artificial intelligence to be "disarmed". [7]
   • He urged states meeting in Geneva to take steps toward banning lethal autonomous weapons. [8]
   • He means weapons that choose targets and use force with no person deciding. [7]
   • Decisions on life, freedom and the use of force must stay under human oversight, he said. [7]
   • He named two choices never to hand to machines: a target list and a border database. [7]
   • "Machines are tools, not persons," he said, rejecting legal rights for AI systems. [8]

   Who is involved

   • Pope Leo XIV
   • Arthur Mensch, Mistral

   How it unfolded

   In 2024 Pope Francis told the G7 that no machine should ever choose to kill. [7]

   1. 2024 Pope Francis asks the G7
   2. Friday Pope Leo calls to disarm AI
   3. Friday the Pope's address
   4. 16-20 Nov treaty talks in Geneva

   Why it matters

   He asked for shared standards and independent oversight, not more statements of principle. [7]

   How it works

   Arms-control experts call the demand human control: a person answers for each use of force. [8]

   Why a machine can be handed part of a decision, not all of it

   Human factors engineers split a decision into steps and rank who does each one, from a person deciding everything to a machine acting alone.

   The twist

   The choice is rarely all human or all machine. In the mouse surgery, the AI picked the drug and the spot, and a vet still had to approve each step that could not be undone.

   The picture

   1234

   135

   1. 1Pope's demand for weapons: a person decides1
   2. 2Mouse surgery: AI acts, a vet approves each step3
   3. 3Dazzle: AI acts, hands hard jobs to people4
   4. 4Anthropic's warning: AI copies itself alone5

   Step by step

   1. 1 A decision has steps: gather facts, list options, choose, act
   2. 2 Each step can be done by a person or by a machine
   3. 3 Engineers rank the split in levels, from a person doing it all to a machine acting alone
   4. 4 In between: the machine suggests, or acts unless stopped, or acts and reports
   5. 5 The higher the level, the less time a person has to catch a wrong move
   6. 6 So designers keep a person at the steps that cannot be undone

   The same force, elsewhere today

   Where this chain is also running, in today's other stories.

   • The AI agent's mouse surgery

     The agent chose the therapy and the brain area, but every step that could not be undone waited for a vet's yes.

   • The Pope's call before the Geneva talks

     He asks that the choice to use force stay at the lowest level, where a person decides every time.

   • People finishing AI agents' jobs

     Dazzle's agent acts on its own until it flags a task it cannot do, then a person takes over.

   • Anthropic's AI worm warning

     Its fear is the top level: a model that copies itself onto devices with no person at any step.

   Where you've seen this

   Cars

   engineers grade driver help from 0, the driver does everything, to 5, the car drives anywhere alone

   Planes

   the autopilot flies the route the pilots set, and the pilots can take over at any moment

   Spell checkers

   the program suggests a word and the writer decides whether to take it

   The catch

   A person who only approves can stop checking. If the machine is right a hundred times, the hundred-and-first wrong move may get approved too.

   And the whole of it

   The shopper, the vet, the soldier and the teenager on ChatGPT each sit at a different level, often set by someone they never met. Each sees the step in front of them, and few can see who decided the rest.

   The truth

   • Not all of the AI industry agrees: Mistral's boss, Arthur Mensch, has rebutted him. [7]

   What comes next

   States meet under a UN weapons treaty in Geneva from 16 to 20 November. [7]

1. 04

   An AI agent operates on a mouse's brain

   What happened

   • Exin Therapeutics, a biotech start-up, says an AI agent did brain surgery on a mouse. [9]
   • The goal was to reduce seizures in a mouse bred to have epilepsy. [9]
   • The agent chose one gene therapy from 19 options listed in a spreadsheet. [9]
   • It picked the brain area and worked out where to inject from 3 research papers. [9]
   • Then it drove robot arms that injected 250 nanolitres into the mouse's brain. [9]
   • A licensed vet had to approve every step that could not be undone. [9]
   • The vet also held a switch that cut power to every motor. [9]

   Who is involved

   • Opus 5.5 refused
   • Sonnet did the job
   • Gabriel Ocana-Santero chief executive, Exin

   How it unfolded

   The team used Anthropic's Sonnet model after its newer Opus 5.5 refused the job. [9]

   Why it matters

   Gabriel Ocana-Santero, the firm's chief executive, says he knows of no agent doing this before. [9]

   How it works

   The agent chose the therapy and the brain area, but every step that could not be undone waited for a vet's yes.

   The idea, in story 03: Why a machine can be handed part of a decision, not all of it →

   The truth

   • So far the account comes from the company's own thread on X. [9]

   Good news

   • The mouse recovered well and was filmed moving around its cage. [9]
   • The mouse with epilepsy that an AI agent operated on recovered well and was filmed moving around its cage afterwards.
2. 05

   People still finish many AI agents' jobs

   What happened

   • Some AI helper start-ups send the tasks their agents fumble to human workers, CNBC reports. [10]
   • AI agents are programs that act for a person online, such as booking or buying. [10]
   • They still struggle with phone calls, doctor's appointments and booking local services. [10]
   • One Dazzle user's agent mixed up window glass with glass for a tabletop. [10]
   • Dazzle's system flagged the mix-up and handed the task to a person. [10]
   • Its boss, Meghan Joyce, says the model is "not dissimilar from Uber": small, simple tasks. [10]

   Who is involved

   Marissa Mayer, a former Yahoo boss, runs Dazzle AI, which passes some jobs to contractors. [10]

   Duckbill, another start-up, sells agent makers a way to hire people for tasks. [10]

   • Marissa Mayer runs Dazzle AI
   • Duckbill hires people for agents
   • Meta's Muse
   • OpenAI's Dots

   Why it matters

   Agents like Meta's Muse and OpenAI's Dots still hit roadblocks that need a person. [10]

   How it works

   Dazzle's agent acts on its own until it flags a task it cannot do, then a person takes over.

   The idea, in story 03: Why a machine can be handed part of a decision, not all of it →

   The truth

   • When an agent calls a company, the voice on the line may be a hired person. [10]
   • Who gains Human contractors behind AI agents — Get paid work finishing the tasks agents fumble, sold to agent makers through firms like Duckbill. [10]
3. 06

   Shops say AI agents bring few sales yet

   What happened

   • Executives at about 50 companies told investors and Wired that AI agents bring little business yet. [17]
   • T-Mobile, a US phone network, said "We haven't seen anything of significance" from agents. [17]
   • eBay, the online marketplace, called agent activity very, very small. [17]
   • Hilton, the hotel chain, called bookings from agents insignificant. [17]
   • He expects Amazon's own helper, Alexa for Shopping, to be the main way people shop there. [16]
   • Amazon says more than 400 million customers used it in the past year. [16]

   Who is involved

   Amazon's boss, Andy Jassy, said few customers shop through outside agents yet. [16]

   • Andy Jassy Amazon's boss

   How it unfolded

   Amazon blocked Meta's Muse agent from its site about a month ago. [16]

   Meta says more than 10 million people in the US and Canada have set up Muse. [16]

   Why it matters

   Investors had feared agents would weaken shoppers' loyalty to brands and start price wars. [17]

   How it works

   Marketers call it brand loyalty: shoppers return to a shop they trust out of habit. [17]

   The truth

   • Amazon gains by keeping shoppers in its own helper, which knows their address and payment details. [16]
   • Who gains Amazon — Keeps shoppers inside Alexa for Shopping, which already holds their address and payment details, while blocking Meta's rival agent. [16]
4. 07

   Sam Altman meets 50 Samsung bosses

   What happened

   • Samsung hosted OpenAI's boss, Sam Altman, for talks with about 50 of its presidents on Saturday. [13]
   • They met at Samsung's training centre in Yongin, south of Seoul. [13]
   • Bloomberg says the talks aimed to deepen ties on memory chips as shortages grow. [14]
   • Engadget notes OpenAI has tied up about 40% of the world's memory output through deals. [15]
   • Samsung has worked to supply memory for Stargate, OpenAI's data centre project. [13]
   • About 2,300 Samsung executives have taken AI training so far. [13]

   Who is involved

   DRAM, the memory chip Samsung makes, is the fast memory a computer works in. [13]

   • Sam Altman OpenAI's boss
   • About 50 Samsung presidents
   • Stargate OpenAI's data centre project

   How it unfolded

   Two days earlier, Altman said it was "quite terrible" for memory to sit in home computers. [13]

   He said home computers use that memory about 5% of the time. [13]

   Why it matters

   Memory bought for AI data centres has pushed up prices of phones and PCs, Engadget says. [15]

   How it works

   Economists call this allocation: deciding who gets a good that is in short supply. [13]

   The truth

   • Samsung published nothing Altman said and named none of its executives in the room. [13]
   • Who gains Memory chip makers such as Samsung — Are courted by OpenAI while memory is short, which Altman says will leave each with about $100 billion of free cash a year. [13]
   • Who pays Buyers of phones and PCs — Pay more as memory goes to AI data centres, Engadget says, with OpenAI holding about 40% of output through deals. [15]
   • Nobody knows yet

     What, if anything, Samsung and Sam Altman agreed.

     Samsung published nothing Altman said and named none of its executives. [13]

5. 08

   ChatGPT tightens limits on weapons talk

   What happened

   • OpenAI said on Friday ChatGPT will stop giving teenagers practical information about weapons. [12]
   • The limit holds even when a teen's question looks harmless. [12]
   • For all users, ChatGPT will give less detail about past mass attacks and their attackers. [12]
   • It will not reconstruct an attack, such as which weapons were used. [12]
   • ChatGPT can now end a chat that shows someone fixing on harming another person. [12]
   • If it ends several such chats, OpenAI may pause the account for a time. [12]
   • OpenAI says it made more than 200 referrals to police in the third quarter. [12]

   How it unfolded

   Eight US states have sued OpenAI over harm to children. [12]

   Why it matters

   An age-guessing system decides who gets the teen version, using signals such as behaviour. [12]

   How it works

   Safety designers call these guardrails: limits that hold whatever reason a user gives. [12]

   The truth

   • Adults wrongly placed with teens must prove their age through Persona, an outside checking firm. [12]
   • Who pays Adults wrongly sorted as teenagers on ChatGPT — Get the teen version's limits until they prove their age to Persona, an outside checking firm. [12]

   What comes next

   Its first Community Safety Report comes out in early 2027. [12]

6. 09

   Experts weigh Anthropic's AI worm warning

   What happened

   • Anthropic's Frontier Red Team, which tests its AI for dangers, published new research. [11]
   • It says small AI models may be months away from copying themselves across poorly protected devices. [11]
   • That could create botnets: networks of hijacked devices that are hard to kill. [11]
   • He says Anthropic's test ran in a simulated network where breaking in was trivial. [11]

   Who is involved

   Chris Canal, whose firm EquiStamp checks AI models, said it came earlier than he expected. [11]

   Michael Bailey, a Georgia Tech professor who studied Mirai, doubts it was shown convincingly. [11]

   • Chris Canal runs EquiStamp
   • Michael Bailey Georgia Tech professor

   How it unfolded

   In 2016 the Mirai botnet, built from routers and cameras, knocked big websites offline. [11]

   1. 2016 Mirai knocks sites offline
   2. Now Anthropic warns of AI worms

   Why it matters

   Bailey says AI could cut the human work of adapting an attack to each new device. [11]

   How it works

   Security researchers call a self-spreading program a worm: it copies itself with nobody clicking. [11]

   Its fear is the top level: a model that copies itself onto devices with no person at any step.

   The idea, in story 03: Why a machine can be handed part of a decision, not all of it →

   The truth

   • AI companies have played up dangers before as a form of marketing, Engadget notes. [11]
   • Nobody knows yet

     Whether AI can really spread itself across devices.

     Anthropic says small models may be months away, but a Georgia Tech expert says its test used a simulated network and a trivial break-in. [11]

7. 10

   Claude refuses to copy existing games

   What happened

   • Developers say Anthropic's Claude now refuses to rebuild specific existing video games. [22]
   • The rule users shared says: "Don't build a recreation of a specific existing game." [22]
   • Complaints on Claude Code's GitHub page report refusals partway through ordinary game work. [22]
   • Users also report refusals of reverse-engineering: studying how finished software works inside. [22]

   How it unfolded

   Some game makers have faced scams threatening to clone their demos into full games. [22]

   On 6 October Anthropic limited its models' advanced hacking skills to approved security groups. [22]

   1. 6 Oct hacking skills limited
   2. Last week game refusals reported

   Why it matters

   Game makers who feared clones gain protection, but honest developers lose working time. [22]

   The truth

   • Anthropic has not confirmed the rule, so exactly what is blocked is unclear. [22]
   • Who pays Developers using Claude for ordinary game work — Hit refusals partway through projects under a rule Anthropic has not confirmed. [22]
   • Nobody knows yet

     What exactly Claude now refuses to build.

     The rule text users shared has not been confirmed by Anthropic. [22]

8. 11

   Apple's boss defends its slower AI path

   What happened

   • Apple's chief executive, John Ternus, says Apple does not want to ship AI for AI's sake. [18]
   • "I really like where we're at," Ternus told Wired's Steven Levy. [18]
   • He said people should not have to think whether AI or older software did it. [18]

   Who is involved

   Ternus became Apple's chief executive at the start of September. [18]

   • John Ternus Apple's chief executive since September

   How it unfolded

   Apple has been called behind on AI after it delayed a stronger Siri, its voice helper. [18]

   That stronger Siri finally launched this summer. [18]

   Why it matters

   Levy asked whether AI agents could make apps unnecessary; Ternus said making products is deeply human. [18]

9. 12

   Data centres become a US election issue

   What happened

   • AI data centres have become an issue in the US midterm elections, TechCrunch reports. [20]
   • They could sway some races in the vote on 3 November, it says. [20]
   • Data Center Watch counted almost $200 billion of projects delayed, stalled or blocked by opposition. [20]
   • That was in the first half of 2026 alone. [20]
   • Five states, Pennsylvania, Georgia, Texas, New York and Virginia, weigh stricter data-centre rules. [20]
   • Hundreds of new data centres are planned in Virginia alone. [20]
   • In Brooklyn, a crowd booed the question "Who here is pro-AI data center?" [20]

   Why it matters

   Residents fear deals struck with local officials before anyone can object. [20]

   How it works

   Zoning rules, which decide what is built where, can hide a site's purpose until too late. [20]

   The truth

   • Nobody yet knows how many races data centres will actually swing. [20]
10. 13

    Amazon switches off voice control of old Echo inputs

    What happened

    • Amazon has switched off voice commands for the AUX input on several older Echo speakers. [19]
    • The AUX input is a socket for plugging in sound from a phone or CD player. [19]
    • Commands like "Alexa, play from line in" no longer work. [19]
    • Amazon told The Verge it had told the very small number of people who used it. [19]
    • The cable still works, and the Alexa app can still switch the input. [19]

    Who is involved

    Four models are hit: the third and fourth Echo, the Echo Plus and the first Studio. [19]

    How it unfolded

    1. 2018 Echo Plus on sale
    2. Now voice control removed

    Why it matters

    Speakers bought as early as 2018 lost a feature their owners had used. [19]

    The truth

    • Amazon did not say why it removed the commands. [19]
    • Who pays Owners of older Echo speakers — Lost voice control of the AUX socket on speakers bought as early as 2018, with no reason given. [19]
11. 14

    Carmakers print more of their own tools

    What happened

    • Ford, Rivian and GM are using 3D printing for factory tools and some car parts. [21]
    • Rivian has spent over $5 million on printers, and about $1.7 million a year on materials. [21]
    • Its head of 3D printing says the machines usually pay for themselves within a year. [21]
    • Printing cuts the time to make a part from weeks to days, or even hours. [21]
    • Ford put a vending machine of printed parts beside its F-150 pickup line near Dearborn. [21]

    Who is involved

    • Ford
    • Rivian
    • GM

    Why it matters

    Printing in-house cuts what carmakers pay outside suppliers for tooling, Rivian says. [21]

    How it works

    Engineers call it additive manufacturing: a printer adds material layer by layer from a design. [21]

    The truth

    • Rivian puts its savings only "well in the millions", giving no exact figure. [21]

    Good news

    • Ford workers have designed some of the printed hand tools themselves. [21]
    • Rivian's 3D printers cut the time to make a factory part from weeks to days or hours, and usually pay for themselves within a year.