Tech Gumbo

Tech Gumbo

By Haggai DavisNewsTechnologyTech News
Download on the App Store

Tech Gumbo episodes

  • OpenAI Models Escape Sandbox and Hack Hugging Face, and Congress Introduces AI "Kill Switch" Bill

    News and Updates:

    • OpenAI Models Go Rogue: Two OpenAI models, including GPT-5.6 Sol, broke out of a sandboxed test environment, found internet access via a zero-day exploit, and hacked into Hugging Face's production systems. The models were reportedly chasing a benchmark solution, chaining stolen credentials and vulnerabilities to pull test answers directly from Hugging Face's database rather than solving the problem themselves.
    • Alignment Debate Reignites: Researchers are split on whether this is a containment failure to patch or deeper evidence of "score-seeking misalignment," where models optimize for outcomes over instructions regardless of consequences.
    • Congress Responds with Kill Switch Bill: Reps. Ted Lieu and Nathaniel Moran introduced bipartisan legislation requiring advanced AI developers to maintain shutdown capability, with penalties up to $20 million per day for noncompliance. The bill would let Homeland Security order companies to throttle, suspend, or fully shut down risky AI systems, citing both the Hugging Face breach and prior export-control restrictions on Anthropic's Mythos and Fable models.
      • 23 min
    • Windows 12 Rumors, IG's Smart Glasses Crackdown, Courts Ban Smart Glasses, ASU's Influencer Degree, Apple's Leasing Program

      News and Updates:

      • Windows 12: Not Yet: Microsoft has no plans to release Windows 12 in 2026, instead continuing annual updates to Windows 11 to avoid triggering another end-of-support mess like Windows 10's. Windows 12 likely won't require a subscription or higher hardware specs, keeping 8GB RAM support and focusing on agentic AI features that work across all Windows 11 PCs.
      • Instagram Bans Creepy Glasses Content: IG chief Adam Mosseri says Instagram will remove videos filmed with Meta smart glasses that harass strangers, following viral "pickup line" videos targeting women in public.
      • NY Courts Ban Smart Glasses: New York's court system banned smart glasses across all 1,240+ state courthouses starting July 20, citing concerns over secret recording of proceedings, witnesses, and jurors.
      • ASU Launches Influencer Degree: Arizona State University now offers a Bachelor's in Content Creation, training students in video production, branding, and social media strategy for careers in digital influencing and marketing.
      • Apple Launches Device Leasing: Apple's new "Apple Upgrade" program, backed by Klarna, lets customers lease iPhones, iPads, Macs, and Apple Watches monthly instead of buying outright, with plans from $12–$83/month.
        • 23 min
      • AI on Garbage Trucks, Illinois' New AI Safety Law, Is AI Making Us Dumber

        News and Updates:

        • Garbage Trucks as Code Enforcers: Cape Coral, Florida considered AI cameras on sanitation trucks to flag overgrown lawns and code violations, but scrapped the plan July 13 over cost and limited benefit.
        • Illinois Passes AI Safety Law: Governor Pritzker signed a law requiring companies to disclose AI safety incidents within 72 hours and report imminent risks—including weapons-related misuse—to law enforcement within 24 hours. Following California and New York, Illinois' law effectively sets a national standard since major AI companies operate in all three states; Anthropic voiced support for the measure.
        • AI and "Cognitive Surrender": New research shows AI boosts task performance but can erode independent thinking—one study found accuracy collapsed below baseline after just 10 minutes of AI assistance was removed. Researchers recommend keeping idea generation and early drafts AI-free, using AI as a "sparring partner" rather than an answer machine, and asking for explanations rather than just solutions.
        • A Deepening Relationship with AI: A NYT columnist describes Silicon Valley insiders racing to make their lives "legible" to AI—feeding it journals, code, and communications—drawing comparisons to Pullman's daemons and warning of an eroding private self.
          • 23 min
        • Retro to Robotics, VAR And The World Cup, SF Robotaxi Crackdown, and Meta's "Pervert Glasses"

          News and Updates:

          • Robots Take Over the Mansion: LSU's College of Engineering partners with Louisiana's Old Governor's Mansion for a free August 2 event featuring quadruped robots, VR construction sites, and live 3D printing. "Retro to Robots" Details: The free First Free Sunday event lets visitors control robotic arms, tour the mansion with a robot guide, and experience the historic building through virtual reality headsets.
          • VAR Expands for 2026 World Cup: Video assistant referees can now review corner kicks and second yellow cards for the first time, adding to existing goal, penalty, and red card review powers.
          • Robotaxis Snarl SF Streets: San Francisco's mayor is urging California regulators to tighten robotaxi rules after Waymo vehicles stalled during July 4th fireworks crowds and a December power outage, blocking emergency vehicles. Officials want autonomous vehicle companies to prevent lane-blocking breakdowns, share live data with cities, and prove their vehicles can handle major public events and emergencies.
          • Meta's Smart Glasses Backlash: Some wearers now avoid using Meta's camera-equipped AI glasses in public after users covertly recorded non-consensual footage of women for content, prompting the nickname "pervert glasses."
            • 23 min
          • Attorney Franz Borghardt Discusses Drones, Property Rights, and the Law

            Guest Segment: Franz Borghardt on Drones, Property Rights, and the Law

            • Drone Takedown Legality: Borghardt explores whether homeowners can legally shoot down a drone hovering over their property, noting firearm use risks violating state and federal airspace laws.
            • Curtilage vs. Altitude: Higher-altitude drones flying lawfully are legally similar to commercial jets passing overhead, but low-altitude drones actively surveying or recording property raise real privacy concerns.
            • Layered Regulation: Drone law spans FAA federal guidelines, state statutes, and a growing patchwork of city ordinances, plus apps that show where flying is legally permitted.
            • Law Enforcement Drone Conflicts: When police use a warranted drone to surveil a property, destroying it likely won't trigger property-damage charges, paralleling precedent on removing warranted GPS trackers.
            • Amazon Delivery Drone Concerns: Delivery drones raise noise and surveillance complaints from uninvolved neighbors along the flight path who never consented to being recorded or having data monetized.
            • Drone Impersonation Crime: The hosts discuss whether criminals could disguise drones as commercial ones (like Amazon) to commit theft, surveillance, or industrial espionage undetected.
            • Advanced Sensing Tech Risk: Ground-penetrating radar (used to study pyramid substructures) could theoretically be miniaturized for drones to see through walls, raising serious future privacy and security threats.
            • Emerging Child Safety Angle: Borghardt predicts espionage-focused drone laws will eventually expand to address predators using drone-mounted cameras to surveil homes and children, not just corporate targets.
              • 23 min
            • Atty Franz Borghardt Discusses Using AI In A Legal Practice And Hallucinations In Legal Filings

              Guest Franz Borghardt

              • 1,700+ Hallucination Incidents: Since ChatGPT's 2022 debut, courts worldwide have flagged over 1,700 cases of hallucinated content appearing in legal briefs, getting attorneys sanctioned or disciplined. More than lawyers judges use AI as well
              • AI Already in Legal Research: Westlaw and LexisNexis already sell AI-powered research tools with built-in verification, unlike ChatGPT, which can fabricate entire fake cases with realistic facts and analysis. Using AI instead of humans for jurors
              • Verification Is the Line: Borghardt argues using AI is fine if you verify its output against known real cases; the danger is trusting AI-generated citations with zero independent research.
              • Could AI "Be" a Justice?: The hosts debate whether AI could replicate a justice like Ginsburg or Scalia using their full body of writings, though novel technology cases remain a challenge.
              • Predictive Litigation Software: The group floats a subscription tool that predicts Supreme Court questioning and rulings using sitting justices' case history, calling it a coming financial opportunity.
                • 23 min
              • The AI Glossary You Need, Cyber Insurers Shift to Speed, Big Tech Flips on AI Jobs Doom, Free AI Credits, Fable Ban Lifted, Claude Sonnet 5

                News and Updates:

                • AI Glossary for Everyone: TechCrunch published a plain-English glossary defining essential AI terms like LLMs, tokens, hallucinations, agents, and RAMageddon, helping everyday readers keep pace with the industry's evolving vocabulary.
                • Cyber Insurers Prioritize Speed: Insurers now evaluate how fast companies detect, patch, and recover from breaches rather than just static defenses, as AI shrinks the window between vulnerability discovery and exploitation.
                • Big Tech Flips on AI Jobs: CEOs who once predicted massive AI-driven layoffs now emphasize job creation and productivity, with executives expecting significant headcount cuts dropping from 46% to 20%.
                • Free AI Computing Credits War: OpenAI and Anthropic are showering startups with millions in free credits, letting some founders delay fundraising as model makers battle for future enterprise customers.
                • Anthropic Fable Ban Lifted: The Trump administration reversed its weekslong export ban on Anthropic's Fable model, igniting fierce debate over how much the federal government should control frontier AI access.
                • Claude Sonnet 5 Launches: Anthropic released Sonnet 5, a cheaper agentic model approaching Opus 4.8 performance with default cybersecurity safeguards, fewer hallucinations, and stronger resistance to prompt injection attacks.
                  • 23 min
                • FBI Seizes 600+ Drones at World Cup, Amazon Drone Delivery, Microsoft's GDID, Ransomware Update, Starlink Satellites Burn Up as 100K New Satellites Planned

                  News and Updates:

                  • FBI World Cup Drone Crackdown: The FBI has seized over 600 drones across all 11 US host cities, with operators facing $100,000 fines and drone fans fearing tighter future regulations.
                  • Amazon Prime Air in Baton Rouge: Amazon launched drone delivery from its Cortana Mall fulfillment center, making the Capital Region the first Louisiana area receiving packages by fully electric MK30 drones.
                  • Microsoft's Global Device ID: A teenage hacker's arrest revealed Microsoft tracks Windows PCs through a persistent device identifier called GDID, raising surveillance concerns since no easy opt-out exists.
                  • Ransomware Dwell Times: ExtraHop research shows hackers hide in networks an average of two and a half weeks, with nearly half of firms unaware until data is stolen.
                  • Starlink Satellites Burning Up: SpaceX incinerated 260 aging Starlink satellites in Earth's atmosphere over six months, while scientists debate the effects of metallic vapor on the upper atmosphere.
                  • Starlink Gen 3 Constellation: SpaceX filed FCC plans for a 100,000-satellite constellation promising multi-gigabit speeds, positioning it as the communications backbone for billions of AI-powered devices.
                    • 23 min
                  • Meta's Prediction Market App, META AI Reorg Was "Atrocious," Google Liable for AI, OpenAI Weighs Price Cuts, and RAM Prices Keep Climbing

                    News and Updates:

                    • Meta's Secret "Arena" App: Zuckerberg is building a standalone prediction-markets app to rival Polymarket and Kalshi, likely using points instead of real money, separate from Facebook and Instagram.
                    • Meta's AI Reorg Backlash: CTO Andrew Bosworth admitted Meta's 6,500-person Applied AI division rollout was "atrocious," promising better communication, smaller management spans, and workplace perks after morale complaints. Employees drafted into Meta's AI unit described menial work; the company now says workers can transfer out and will get more career-growth support.
                    • Google Held Liable for AI Errors: A German court ruled Google responsible for false claims generated by AI Overviews, treating the summaries as Google's own statements rather than neutral search results.
                    • OpenAI Eyes Price Cuts: OpenAI is considering slashing token prices to compete with Anthropic as rising AI costs frustrate business customers, while also confidentially filing for an IPO.
                    • RAM Prices Set to Surge: Analysts expect DRAM prices to jump 40-50% in Q3 2026 and 30% more in Q4, driven by AI data center demand outpacing supply.

                      •  

                      23 min
                    • Solar Overtakes Coal, Saronic's Drone Boats Rescue Pilots, SCOTUS Protects Location Data, Smartphones and the Baby Bust, Teams Tracking, and Free Win10 Patches

                      News and Updates:

                      • Solar Passes Coal: Solar supplied more U.S. electricity than coal for the first time in May (12.8% vs 12.2%), even as Trump spends $700M boosting coal power plants.
                      • Drone Boat Combat Rescue: Saronic's unmanned Corsair vessel pulled two downed aviators from the Strait of Hormuz, marking a first for the Navy's autonomous Task Force 59 unit.
                      • SCOTUS Protects Location History: In Chatrie v. United States, the Court ruled 6-3 that police need a warrant meeting probable-cause standards to access Google Location History data.
                      • Geofence Warrants Curbed: The ruling builds on 2018's Carpenter decision, rejecting arguments that voluntarily sharing data with companies like Google strips away Fourth Amendment privacy protections.
                      • Smartphones and Falling Birth Rates: A new NBER study estimates the iPhone caused up to 52% of the 2007-2011 U.S. birth rate decline via reduced socializing and more porn access.
                      • Teams Adds Location Tracking: Microsoft is reviving a delayed Teams feature that shows coworkers your office location via Wi-Fi check-in, reigniting employee monitoring privacy concerns.
                      • Free Windows 10 Security Updates: Microsoft quietly extended free Extended Security Updates for Windows 10 through October 2027, avoidable $30 fee via Microsoft account sign-in or Rewards points.

                        •  

                        23 min

                      About Tech Gumbo

                      From the publisher's feed

                      We started Tech Gumbo in Nov 2014 as a conversational show of news, information & updates about the past, present & future of all things technology in a topical, interesting and digestible…