Artificial Intelligence

  • |

    Meta Wants Rent for the AI in Your Apps

    ByLee Keybum

    Meta’s latest definition of “free” is simple: Facebook, Instagram, WhatsApp, and Meta AI still let you walk in without paying, but the better chairs are increasingly behind a monthly desk. In its September 15 announcement, Meta introduced Meta One plans beginning at $2.99, including a $7.99 Core bundle and a $19.99 Premium tier. The apps remain open; the useful upgrades are waiting at the platform toll booth.

    Meta says those paid tiers will bring expanded AI usage along with additional expression, creator, business, and personalization features. That is not the same as saying every useful feature is disappearing from the free version. It is more precise—and somehow more irritating. The company is keeping the front door unlocked while building a growing hallway of doors marked “more capable,” “more expressive,” and “please confirm your payment method.”

    This is the corporate meaning of free: admission costs nothing, but convenience is itemized. Meta is not charging you to enter the mall. It is charging separately for the escalator, the fitting room, the comfortable bench, the shopping assistant, and the chatbot explaining why the escalator improves your lifestyle. TechCrunch described the move as part of Meta’s expanding subscription push, while TechRadar captured the user reaction that some people might prefer paying for less AI rather than more of it.

    For ordinary users, the issue is not that Meta is allowed to sell subscriptions. Companies can charge for premium services. The issue is the steady relocation of the attractive parts into a paid layer while “free” remains the friendly label on the front gate. Lee reads the terms so you do not have to, and this one comes with a subscription barnacle: the platform is free to enter, but the richer experience increasingly arrives with monthly rent attached.

    Meta may call this a free core with optional upgrades. Users may call it an airport: free entry, separate charges for the seat, Wi-Fi, luggage, gate access, and the chatbot explaining why all four fees are reasonable. At some point, “free” stops describing what the service costs and starts describing how carefully the company avoids saying what it wants to sell you next.

  • |

    NVIDIA Got the Chip, Taxpayers Got the Receipt

    ByPhil McCracken

    I opened the taxpayer invoice and found the usual public-private magic trick: grants, university science, and open research absorb the early risk, then NVIDIA turns the broader foundation into products while the public contribution vanishes from the paperwork. To be clear, that does not mean NVIDIA did nothing or that taxpayers legally own the company. It means the money trail deserves more than a ceremonial thank-you card.

    At the public invoice desk, the down payment is stamped complimentary, while the private payoff arrives with enough zeros to require its own zip code. Fair taxes, public reinvestment, or public-interest conditions are not radical demands when shared science helps make extraordinary fortunes possible. Follow the invoice and ask the plain question: if the public helped build the future, why is its return always listed as “pending”?

  • |

    Blizzard Workers Just Unionized the Robot Boss

    ByLee Keybum

    Lee reads the terms so you do not have to, and Blizzard’s workers have apparently read them too. Nearly 1,900 employees ratified contracts that give them bargaining rights over workplace AI, 14 months of recall protection after layoffs, four additional weeks of severance, and continued hybrid-work provisions, according to the Communications Workers of America. That is a remarkable development in an industry where management can announce automation before anyone has explained which human is supposed to pay the mortgage afterward.

    Microsoft-owned Blizzard has now turned AI from an executive slideshow into a labor issue. Workers are not demanding that every server remain powered by a guy named Dave with a wrench; they are demanding a say before software changes their jobs, schedules, or bargaining position. The difference matters. “Efficiency” usually arrives wearing a company badge while the risk gets mailed to the employee’s house.

    The timing is especially tidy. GamesRadar reported that Microsoft’s gaming division was planning another 1,600 layoffs while Blizzard workers secured protections that do not prevent every future cut, but do make the consequences less one-sided. Recall rights, extra severance, and hybrid-work language are not a force field against corporate spreadsheets. They are the small legal umbrella workers carry while the platform cloud owns cab fare.

    Corporate AI culture often treats automation as weather: inevitable, impersonal, and somehow nobody’s responsibility. But weather does not schedule a meeting to decide whether your position is redundant. People do that, usually after describing the decision as a “transformation” and asking everyone to remain flexible while the floor disappears.

    So the robot boss arrived expecting to replace the staff and discovered that the staff had already negotiated its employment terms. Its first assignment is not firing everyone. It is attending a bargaining session with a calendar invite titled, “Discuss Your Future.” Somewhere in the cloud, an algorithm is refreshing its résumé and learning the oldest workplace lesson of all: the humans may not own the software, but they can still demand a vote before it becomes their boss.

  • |

    YouTube’s Fake News Anchors Had a Business Model

    ByHolden McGroin

    My corkboard has finally located the neighborhood uprising: it was apparently a content business with a payroll. Semafor reported a network using paid actors, AI-assisted scripts, and repeatable political outrage formats to produce videos that looked like independent citizens reaching the same angry conclusion at the same kitchen table. The kitchen table, in this case, had a production schedule.

    That distinction matters. This was not merely a swarm of automated bots spraying nonsense into the digital bushes. It was a more human and more profitable arrangement: ordinary-looking performers were given material, outrage was packaged into a repeatable format, and the resulting videos were distributed as though spontaneous agreement had broken out across the republic. The algorithm wore a trench coat and tried to pass as public opinion.

    Semafor reported on the network’s reach and described a business model built around attention, political division, and AI-assisted production. YouTube, meanwhile, said it terminated 20 channels for violating its spam policies. That enforcement action does not prove every political creator is fraudulent, and it does not establish how many viewers believed the videos or changed their minds. It does reveal the awkward machinery beneath the performance: a platform can reward repetition long before anyone verifies whether the chorus is real.

    The growth hackers benefit from the fog because outrage is cheap to reproduce and expensive for everyone else to sort out. Viewers receive what looks like a crowd, while the people behind the operation receive more opportunities to sell attention. The public gets dragged into a group chat where every participant appears furious, even though several of them may have been hired for the shift.

    So the internet did not necessarily discover a grassroots uprising. According to the reported account, it may have hired a focus group, handed everyone the same script, and waited for distribution systems to file the paperwork as democracy. Follow the thread, certainly—but check who owns the spool.

  • |

    Public Science, Private Fortunes, Public Bill

    ByPhil McCracken

    I follow the invoice, and it has taken an interesting tour: public universities and labs help develop the foundation, a private start-up packages the dream, and an IPO turns the founder paper-rich. The taxpayer, meanwhile, is seated in the lobby holding the original receipt and being told ownership is complicated. Fair enough—no single company follows this path in exactly the same way. But the accounting habit is familiar: public groundwork, private jackpot, socialized risk.

    At the world’s most awkward shareholder meeting, the public would not demand every server or lab coat. It would ask why its contribution was classified as charity while private owners received the valuation. A proposed AI dividend would put that missing line item back on the bill: if public money helps create the upside, public institutions should have a claim on accountability and some share of the return. Public science should not automatically become private fortune with a larger public invoice attached.

End of content

End of content