
In this episode, Ray Cochrane leads with Meta, Google, and TikTok suing California to block a law that would remove personalized feeds and autoplay for kids under 16. Next, he covers new top AI models from Anthropic and Google, along with the price range for Anthropic’s IPO. He also gets to Jensen Huang’s resignation from Nvidia, a long Microsoft 365 outage, and Florida’s worst dengue outbreak in years.
– Want to start a podcast? It’s easy to get started! Sign up at Blubrry
– Thinking of buying a Starlink? Use my link to support the show.
Subscribe to the Newsletter.
Email Ray if you want to get in touch!
Like and Follow Geek News Central’s Facebook Page.
Full Summary
Ray Cochrane opens with a studio update. The rewiring is finally done, and everything has been cleaned up and reorganized. Now the TriCaster and the sound gear sit within easy reach. After weeks of audio trouble, he has found other ways to make things work.
On the personal side, Cochrane is heading to Detroit on Saturday for some time at a museum. Meanwhile, rain and cold have rolled in. So he reminds listeners to keep warm, and socks help.
Big Tech Sues to Stop California’s Teen Feed Law
The lead story comes from KQED. On November 6, the public radio station reported that Meta, Google and TikTok are suing California. Each company filed its own federal lawsuit on November 4, and YouTube joined Google’s. They want a judge to block AB 1709 before January 1. On that day, kids under 16 in California would lose personalized feeds and autoplay.
The law doesn’t lock kids out of the apps. They can still search, send private messages, and see posts they asked for, such as videos from creators they follow. However, a platform that won’t offer that stripped-down version has to delete the accounts and the data tied to them. Parents can’t turn the features back on, either.
The bill’s author, Assemblymember Josh Lowenthal, told NBC Los Angeles that “certain public health issues trump parental controls, and this is absolutely one of them.” On the other side, Google spokesperson José Castañeda told CalMatters, “We believe parents are best positioned to make the right decisions for their families.” Cochrane sides with Google on that point.
Platforms also have to check a user’s age before serving any personalized feed. A separate state law has phones and app stores pass apps an age range, starting the same day. Only the attorney general and local prosecutors can enforce AB 1709. Fines reach $50,000 per affected kid for knowing violations, or $25,000 for careless ones. The bill passed the state Senate 39 to 0, and Governor Gavin Newsom signed it on September 10.
Free Speech and Section 230
All three lawsuits name Attorney General Rob Bonta. Meta and Google filed in Northern California, while TikTok filed in Los Angeles. Their two main arguments are the First Amendment and Section 230. Additionally, TikTok says some terms are too vague, including autoplay, since the law never defines it.
The free speech case leans on the Supreme Court’s 2024 ruling in Moody v. NetChoice. There, the court said a platform choosing and ordering posts is its own expression, much like an editor laying out a front page. Still, it left open feeds that simply follow a user’s clicks. Meta’s complaint says the law would “appoint the State as the editor, curator, and controller of online speech.” Cochrane calls that a mild digital extreme meant to make the point.
Section 230 is a 1996 federal law that mostly shields websites from blame for what users post. The companies argue a feed only arranges user posts, so the state can’t punish it. California disagrees. A state Senate committee analysis says the law targets how apps are built to keep people hooked, not what users post.
Lowenthal told NBC Los Angeles, “We’ve expected lawsuits all along. This is straight out of the big tech playbook.” Both sides also cite research on teens. Studies link heavy social media use to anxiety, depression and poor sleep. Yet an Assembly committee analysis notes that cause and effect is hard to prove, since nobody can ethically run that experiment on kids.
California Has Been Here Before
In 2024, the state passed SB 976, which limits personalized feeds for everyone under 18 unless a parent consents. The industry sued over that law too. On August 5, however, federal judge Edward Davila refused to put its feed rules on hold. He called that kind of ranking “merely a mirror that reflects back to users their own perceived interests.” The companies have appealed, and the state wants the new cases sent to Davila as well.
Meanwhile, Meta already agreed to teen limits in a separate case. On August 26, it settled with attorneys general from most states, California included. The state puts the deal at up to $17 billion over 10 years. Under it, Facebook and Instagram are capped at two hours a day for users under 18, the apps go dark from midnight to 6 a.m., and teens can turn off autoplay.
Meta’s complaint holds those protections up as gentler ways to reach the state’s goals. As PPC Land pointed out, though, the complaint never mentions the settlement. Cochrane finds that very convenient for Meta, given the money involved.
Facebook’s New Your Algorithm Page
The law still lets kids see posts they clearly asked for, but Meta reads that exception narrowly. Following an account might count, while wanting more baseball might not. PPC Land applied that reading to Facebook’s newest feed tool. By Meta’s own logic, it said, teens under 16 couldn’t use it either.
On October 13, Facebook started rolling out a page called Your Algorithm. It shows a one-line summary of what someone has been watching, followed by lists of topics they want more of and less of. Users can add, move, or delete topics. Then, on November 10, Meta engineers explained that an AI model builds the starting list from past activity. In their words, “A new interest should show up, not take over.”
Meta hasn’t shared usage numbers. Techlicious also noted that Meta didn’t say where typed topics are stored, how long they’re kept, or whether they help train its AI.
A Tight Calendar
TikTok’s hearing is set for December 14 in Los Angeles. Meanwhile, Meta, Google and the state have asked for a joint hearing on December 17. Both companies want a ruling before January 1, and they may seek an emergency order if the hearing slips.
Cochrane wonders whether the companies are quietly building compliance in the background, in case the law holds. He recalls an earlier debate about how a calculator app could verify someone’s age. Thankfully, that job has shifted to the app stores. In the end, he agrees feeds can be addictive and that kids can’t really consent to that, so he’ll keep following the case.
Sponsor: GoDaddy
Economy hosting $6.99/month, WordPress hosting $12.99/month, domains $11.99. Website builder trial available. Use codes at geeknewscentral.com/godaddy to support the show.
Anthropic and Google Ship New Top AI Models
Two of the biggest AI labs released new top models within days of each other. On November 9, Anthropic launched Claude Fable 5.2 alongside Claude Mythos 5.2. Anthropic says they’re one model underneath. Fable carries safety limits on cybersecurity and biology, and any paying customer can use it. Mythos drops those limits, so it only goes to organizations Anthropic has vetted.
Cochrane thinks that barrier is a good thing. In his own security research, lower-tier models often flag his conversations, so he hopes the safeguards keep improving. Anthropic says the new release runs up to 30 percent faster than 5.1 at a 20 percent lower price.
Google’s release is a whole new major version. Gemini 4 arrives as three models named after chemical elements, and episode 1880 covered the first, Argon. On November 10, Carbon reached developers and AI Ultra subscribers. Then today, Google released Hydrogen, a faster and cheaper model now live in Google Search worldwide. Cochrane warns listeners to be careful with those AI answers, since the models can hallucinate.
Artificial Analysis now ranks Fable 5.2 first overall at 65, with Carbon at 60. It calls both “notably slow per task and very verbose.” Google also admits a weak spot in Carbon: prompt injection, where hidden instructions in a page or file trick an AI. Google wrote that “prompt injection robustness remains an area for continued improvement.” Cochrane finds that concerning, pointing to the OpenClaw users whose data leaked early on.
Anthropic Puts a Price Range on Its IPO
This one follows up on episode 1883, when Anthropic filed to go public. On November 10, TechCrunch reported the price range. Anthropic is offering 825 million shares at $70 to $78 each, or about $64 billion at the top. It plans to set the final price on November 18 and start trading the next day.
The company’s value depends on what you count. By Reuters’ math, the shares outstanding after the sale come to about $1.02 trillion. Counting options and convertible notes, it’s closer to $1.4 trillion. Both fall short of the $2 trillion investors were eyeing earlier this year. Meanwhile, Morningstar’s analyst values a share at about $92 and blames the gap on investor mood around AI.
Additionally, seven companies will buy $26 billion in shares at the IPO price: Google, Nvidia, Microsoft, AMD, Broadcom, Micron and TSMC. Cochrane calls that amount crazy. Regular investors also get up to 20 percent of the shares, compared with the more typical 5 to 10 percent. With the IPO a week away, he expects big news despite the shift in public sentiment on AI and data centers.
Jensen Huang Resigns From Nvidia
Episode 1884 led with Jensen Huang, and now he’s out. This morning, Nvidia announced its board had accepted his resignation, effective immediately, without giving a reason. Four days earlier, Huang had told employees, “I am fully committed to NVIDIA and will work to earn back your confidence.”
In between, The New York Times reported a second affair, this time with an executive who worked under him. Neither Nvidia nor Huang has confirmed it. However, a person familiar with the board’s decision told CNBC the mood turned after that report. CFO Colette Kress is now interim CEO and keeps the finance job too.
Analysts at Oppenheimer warned of a possible brain drain, though no other departures have been reported. Nvidia stock fell about 14 percent after the opening bell, then closed down about 4 percent. Earnings still arrive November 17. Additionally, Representative Brian Mast says Huang has agreed to come before Congress, likely before Thanksgiving.
Cochrane calls it poor decision-making, especially with the first affair tied to national security concerns. Still, he finds it somewhat relieving that Nvidia hasn’t seen an exodus.
Microsoft’s Business Email Goes Down for Hours
Microsoft had a major outage today, centered on Exchange Online, the business email service behind Outlook. For many users, mail stopped moving entirely. Microsoft also warned that Teams, SharePoint, and Copilot could be affected. Outlook reports on Downdetector peaked around 130,000 just after 1 p.m. Eastern. As Cochrane notes, that’s only one tracking site.
Microsoft’s explanation kept shifting. First it blamed heavy load, then said the load wasn’t extra traffic. Failing servers were pulled, which piled more work on the rest. At 5 p.m., it said its fixes “have not provided the anticipated relief.” By evening, restored servers were falling over again.
The damage spread further, too. Anthropic and OpenAI said Claude and ChatGPT lost their connections to Microsoft 365 mail and calendars. SecurityWeek noted nothing pointed to hackers. Cochrane jokes that AI-run infrastructure is his guess, which would mean no human can figure out the problem.
Apache’s Web Server Gets a Fix for a Critical Bug
Cochrane ran into this one during his morning browsing. SecurityWeek reported today that Apache’s web server has a fix for a critical flaw in its HTTP/2 handling. With HTTP/2 on, an attacker could run their own code without logging in. The fix is version 2.4.71.
Anthropic’s free OSS Scanner flagged it first, and two Google researchers found it independently. Red Hat says “public research has demonstrated a reliable exploit,” although Wiz reports no attacks so far. Until you can update, turn HTTP/2 off. Cochrane calls it scary and urges server admins to act now.
NIST Writes a Security Guide for One-Person Businesses
On November 2, NIST released a free security guide for one-person businesses like freelancers and gig workers. NIST says more than 28 million US businesses fit. Co-author Daniel Eliot said it “assumes the reader has no familiarity with cybersecurity or the NIST CSF.” It replaces a 2016 guide written for small businesses in general, which Cochrane calls a well-needed update.
The guide opens with nine first steps. Turn on multi-factor authentication, starting with email and banking, and use passphrases of at least 15 characters. Next, keep backups that actually restore, update software, and change factory passwords on routers and printers.
Cochrane adds his own advice: get a reliable password manager, but check its history for leaks first. Finally, in a callback to episode 1884, the guide’s AI section is titled Super Intelligence. It says to list your AI tools, check their privacy settings, and review their output, which he calls common sense.
Scientists Map a Deep Brain Region Cell by Cell
Nature reported today that hundreds of scientists in nine countries have mapped the basal ganglia one cell at a time. That region helps steer movement, habits, mood, and decisions. The teams profiled about 25.5 million cells from humans and two kinds of monkeys, and they published 26 papers today.
Addiction, schizophrenia, bipolar disorder, Huntington’s and Parkinson’s are all linked to the region. Hongkui Zeng of the Allen Institute said, “Disease conditions are always cell-type-specific.” One team also found the striatum splits into six zones visible only through gene activity. With age, those zones blur, most of all where Huntington’s and Parkinson’s do damage.
The cells came from just six human donors, three macaques and two marmosets. Joseph Ecker of the Salk Institute told Reuters that treatments will take years. Next comes the whole brain, and Zeng told The Transmitter, “We hope to finish everything in five years.” Cochrane calls that mind-blowing and hopes to see it.
Dengue Spreads Around Tampa
Smithsonian Magazine ran an explainer today on Florida’s dengue outbreak. Dengue is a virus spread by mosquito bites, not from person to person. Through October 31, Florida counted 524 locally caught cases this year, compared with 62 in all of 2025. By the Associated Press’s reading, no state has had a bigger dengue year since at least the 1940s.
Most cases are around Tampa, and Hillsborough County alone has 407, about four in five. Cochrane had assumed everyone who catches dengue gets a fever. In fact, the CDC says only about one in four feel sick. About one in 20 of those develop severe dengue, and one person has died so far.
Three counties around Tampa Bay declared local emergencies, and the CDC opened an emergency response at its lowest level. There’s no dengue vaccine available in the US. Instead, the CDC recommends repellent, long sleeves, and dumping standing water. Dawn Wesson of Tulane University told the AP, “There’s a good chance that this is the new normal.” Cochrane thinks that may be a stretch, since the outbreak is concentrated in one county.
Cochrane wraps up with housekeeping. Listeners can become a GNC Insider at geeknewscentral.com/insider, email [email protected], and sign up for the newsletter at geeknewscentral.com. He also points to podcastapps.com and mentions he uses Pocket Casts. Finally, he thanks GoDaddy for more than twenty years of support.
He also reveals this was his second recording of the night. The first time, the TriCaster showed it was recording, but the timer sat at zero, and nothing showed up on export. He wishes everyone an awesome weekend and promises to be back on Monday.
The post Feed Fight #1885 appeared first on Geek News Central.
More episodes from "Geek News Central Podcast"



Don't miss an episode of “Geek News Central Podcast” and subscribe to it in the GetPodcast app.








