13 arrested as students occupy OpenAI's DC lobbying office

Share
13 arrested as students occupy OpenAI's DC lobbying office

Washington became the front line of the AI backlash this week: students got arrested inside OpenAI's lobbying office, and the Federal Reserve is openly split over whether the AI buildout is making inflation worse.

More than 30 student activists occupied the front lobby of OpenAI's Washington, DC lobbying office Monday afternoon, and police arrested 13 of them on unlawful-entry charges after a two-hour standoff. The demonstrators — organized by the climate group Sunrise Movement and the anti-AI campaign QuitGPT — unfurled a "Stop Stealing Our Future" banner inside the Gallup Building, where OpenAI keeps a public-facing office, and held signs reading "OpenAI bought my Senator." Metropolitan Police records show officers split the group up for processing: seven to the First District station, four to the Second, and two to the Third. Organizers say the action is one of many across more than 20 states this August under a "Dump Big Tech" banner, and student organizer Raya Gupta framed the risk in generational terms: "My friends chose to risk arrest to bring the fight to OpenAI."

The target is not accidental. OpenAI opened the office in May as its super PACs push more than $140 million into the midterm elections to steer AI regulation — and its policy team argued just this morning that the real rulebook is being written below the federal level: OpenAI says US AI policy is being written in state capitals. Thirteen arrests in a lobby won't dent a lobbying budget, but the action is the visible edge of a wider shift: AI opposition has moved from comment sections into direct action at data-center hearings, agency offices and now lab lobbies, aimed squarely at the money OpenAI is spending to shape the rules it will have to live under.


The Federal Reserve is openly split over whether the AI buildout is inflation before it is productivity — and the price data is landing on the hawkish side. Goldman Sachs Research sees US AI capital spending hitting $581 billion this year, about 1.8 percent of GDP, and the costs are already visible: household electricity prices up 10.1 percent over two years, DRAM prices on track to climb 400 percent versus 2024, and software costs up 22.9 percent since mid-2024. Corporate adoption is the bottleneck — a Census Bureau survey found only 17–20 percent of US businesses using AI, and OpenAI's own chief economist, Ronnie Chatterji, says the usage gap between power users and average firms has widened to eightfold, adding that productivity gains will take "a little while" to show up in the statistics.

That split has become a policy fight. Minneapolis Fed president Neel Kashkari dissented at July's meeting in favor of higher rates, citing data-center demand, and governor Lisa Cook says she is "prepared to act" if disinflation doesn't arrive; chair Kevin Warsh — who has put Marc Andreessen and Stanford's Charles Jones on a task force studying AI's economic effects — insists a supply response will follow and calls the dispute one of the Fed's "good family fights." The stakes for AI are direct: if the inflation-first reading wins, rate hikes would slow the very capex boom, now roughly $1 trillion globally, that is funding the entire buildout.

What to watch: whether Wednesday's CPI report shifts the Fed's internal balance on AI-driven prices.

Thirteen arrests in a lobby won't change a lobbying budget — but will direct action change how AI companies spend on politics? Tell us in the comments.

Read more

Houmo moves compute inside memory — and A20 enters production

Houmo moves compute inside memory — and A20 enters production

Beijing chipmaker Houmo used its partner conference this week to lay out two compute-in-memory routes at once — and to say the first chip built on the newer one is already in production. Houmo's A20, its first chip built on a 3D compute-in-memory architecture, has entered production. The part supports up to 16 chips fully interconnected, claims a 50% improvement in compute utilization over the previous generation, and covers models from 3B to 30B parameters, according to co-founder and product

Kling 4.0 goes live with 30-second native video at 10-bit HDR

Kling 4.0 goes live with 30-second native video at 10-bit HDR

Kuaishou's Kling AI has spent two years becoming the video model people actually finish projects with. Its new flagship, Kling 4.0, went fully live on October 15 — and this round is less about prettier frames than about the two things that make a generated shot usable in a real edit: length and control. Kling 4.0 generates up to 30 seconds in a single native pass, outputs 10-bit HDR at 4K, 1080p and 21:9, and takes up to 15 reference assets in one task. That budget splits across ten images, fi

Brain organoids classified Braille at 83% — in a trio, not alone

Brain organoids classified Braille at 83% — in a trio, not alone

Two stories today: human brain tissue on a chip decoded a tactile signal, and the data-centre company behind some of the largest AI contracts in Europe turns out to be six months behind on its flagship site. Human forebrain organoids grown on a microelectrode array classified Braille letters with 61 percent accuracy on their own and 83 percent when the responses of three organoids were combined, in a peer-reviewed study from the University of Bristol. The team took event-based tactile readings

Suleyman says Microsoft's base model beat OpenAI's on a quarter compute

Suleyman says Microsoft's base model beat OpenAI's on a quarter compute

Microsoft finished pre-training its biggest in-house model yet — then used it to make a point about the company it bankrolls. Microsoft AI completed pre-training of MAI-Base-2, a mixture-of-experts foundation model with 80 billion active and 2.8 trillion total parameters, and says the recipe is 4.2–12.3x more pre-training FLOPs-efficient than DeepSeek-V4-Pro-Base. The run itself is the number worth keeping: 16,000 Nvidia GB200 GPUs for 38 days across August and September, at 21% average model