Anthropic-Small AI Models Could Self-Replicate via IoT Devices; SpaceX Limits Starlink in Ukraine Hampering Military; OpenAI Secret Recorder in Office-AI Helps Finger Perpetrator; Netflix Laying Off 1,000 Employees 

Anthropic has issued a warning that small AI models may be able to replicate themselves soon and burrow into all sorts of Internet of Things devices like smart TVs and smart refrigerators. Arstechnica.com reports that this would be like computer worms, which propagate without human interaction and can infect IoT devices…only on steroids— with the AI being behind the small models. Anthropic also warned that this new type of malware could be difficult to stop. Bill Gates also sounded the alarm about this in the last several days. Smaller AI models require much less powerful hardware to run than do frontier models. I love tech, as is obvious, but for once am glad to have a ‘dumb’ refrigerator!

SpaceX has placed a limit on some Starlink devices in Ukraine, disrupting combat drone missions that need the tech. According to reuters.com, SpaceX never even gave the Ukrainian military a heads up…and some missions had to be cancelled. Apparently, during flights the connection could drop for 10-20 seconds to as low as 66 mph. The drones kept flying, but it made them much harder to control. One has to wonder if this was because of the damage they have been able to cause to Russian data centers, which seems to have enraged Elon Musk. SpaceX has not responded to requests for comment on why the service was throttled. Ukraine drones don’t use Starlink in Russia, where the service is not available…but instead use other means of communication. 

A spy had connected a recorder on the back of a TV in a conference room at one of OpenAI’s San Francisco offices. After the device was found, mashable.com says the AI giant used its own models to “understand the device and where it had come from.” The models helped finger a Saudi national who had worked there as an engineer. He was arrested by the FBI on October 9th. The recorder was plugged into a USB port on the back of the TV and was wrapped in toilet paper. Apparently he had planted a 2nd device in the building and that was found as well. The device was only in the room a handful of days. OpenAI says it currently has no indication that the man…one Abdulrahman Alfozan… successfully obtained model weights, source code, or customer data. The room where the device was planted, however, was a place where “highly confidential” technical, commercial, and strategic information was frequently discussed.

Netflix is laying off about 6% of its workforce. TechCrunch.com reports that Netflix claims most of the cuts would affect its product, tech, legal, and human resources departments. The company said it will “continue to invest aggressively in high-impact areas” to improve its service, and stated it is on track to deliver its full-year forecast. The streamer has had 16,000 full-time employees. This layoff comes in the aftermath of the battle between Netflix and Paramount Skydance over acquiring Warner Brothers. 

I’m Clark Reid and you’re ‘Technified’ for now.


Apple’s Home Event-HomeView & More

After years of rumors, Apple finally unveiled the HomeView. The new device has a 7-inch ‘Liquid Retina’ display, and a domed speaker base. There’s a stainless steel hinge with tilt support. The new HomeView can dim down more than other Apple devices, so could be used as a nightstand unit if you don’t use the iPhone horizontally for a bedroom clock. 

There is a separate wall-mount version of the device, too. It can be mounted by you or you can opt to drop bucks for a pro to do it. The HomeView can be preordered today for $299. The mounting kit for wall mounting will set you back $59.95 for the cheapest, or there are a couple other models that run $79.95 and $129.95. The pricy one is out next year. HomeView can recognize up to 6 users, and according to Apple it will deliver a “personalized experience” for each one. This includes a personalized Home Screen with apps and widgets. “Personal content stays private from others, while shared photo albums and music can be easily enjoyed by everyone in a household, including guests.” 

The gadget supports Siri AI. As far as audio, it is no match for the HomePods. HomeView has 2 full-range drivers, a down-firing woofer, and two force-cancelling passive radiators. The 1st gen HomePod had 6 drivers and a woofer, and 2nd gen one has 5. Note that the wall-mounted unit doesn’t have the good speaker system. The table-top one comes in silver, starlight, rose pink, and space gray. The wall-mounted HomeView comes in just space gray or silver. HomeView isn’t initially available in the EU. That’s because Siri AI is not available in the EU yet. Also, there is only limited support for third-party apps. 

Also announced today was the HomePod Mini 2. The updated Mini ships October 30th, but won’t support Siri AI at launch. It comes in green, pink blue, midnight, and white. The upgrade from the S5 to S10 chip should make the Mini’s capability better…honestly, version 1 is pretty helpless and the sound isn’t great. The HomePod Mini 2 is a pretty big jump in price from the original. It’s $149…compared to $99 for the first model. 

I won’t cover the Apple Home Collection today, as these leaked and were already covered by me before…but they include the new door lock made in conjunction with LG, and the outdoor cams and thermostat. 

I’m Clark Reid and you’re ‘Technified’ for now. 


Global Smartphone Sales Down-Chip Shortage; Cars Sharing Data with Third Parties; HomeView & HomeOS-Likely Apple’s; Parents Angry Over Planned Tesla Fremont Increased Emissions

Counterpointresearch.com reports that the global smartphone market declined 20% year over year in the third quarter of 2026, reaching its lowest third quarter level in 14 years…primarily due to continued memory chip shortages. Samsung continues to top the market with a 23% share, powered by better-than-competitive positioning in securing shipments and robust early demand for its Z8 foldable series, led by the Z Fold 8. Apple’s market share got past the 20% mark for the first time in any third quarter, reaching 22%, despite a year over year decline in shipments. The market saw a widening gap between premium and price-sensitive segments, with premium demand remaining relatively strong while higher device prices put greater pressure on affordability, hurting the lower end phone sales.

You likely won’t be surprised, but still pissed about this. Many new cars are sharing your data with third parties. According to engadget.com, top US car maker General Motors’ app has sent data…including the vehicle ID number to Snap, Yahoo, Meta, Microsoft, Adobe, Google, Pinterest, and data broker Acxiom. This all according to a study from Northwestern University. The GM apps include myCadillac, myChevrolet, myBuick, and myGMC. HondaLink sent your VIN and location to analytics firm Amplitude. Lincoln and myNissan apps also were guilty of sending off your data to third parties. Because a VIN can’t be changed, the third parties can pair that with your email, and they can tie you to activity on other apps and websites. To give credit where it is due, Honda has asked Amplitude to delete location data, and then updated the HondaLink app to stop sending it. 

Apple is going to hold their ‘Welcome Home’ event tomorrow, rolling out their new smart home hub and other goodies. We will cover it of course, but right now a little tidbit or two besides the rumored devices has leaked out. Macrumors.com notes that a company called Home Services LLC secretly filed for a trademark for “HomeView” in the US and other countries. The firm has been tied to Apple. Another Apple-connected company filed for HomeOS. Speculation is that the new home hub might be called HomeView. At any rate, we’ll know tomorrow…and will see it as well as several other either new or updated Apple hardware.

There is an elementary school near the Fremont, CA Tesla factory. It’s very close…less than 1,000 feet away. Thenextweb.com says parents of school kids want more time to question a plan that could raise emissions from the factory. Tesla has applied to change emission limits on two aluminum melting furnaces and add new equipment to cut nitrogen oxides, according to a notice from the Bay Area Air District. Notice of the proposed change came over one single message on a school app on the third of September. That notice said that Limits would be modified…but didn’t specify that they would go up. Then, someone posted on an anonymous Substack that the school district’s own engineering papers showed that emissions would go up 13-fold! Parts of the engineering review were blacked out, too. The parents have a petition that has gathered over 1,000 signatures demanding more time for the comment period. Tesla has not replied to requests for comment from the EV maker.

I’m Clark Reid and you’re ‘Technified’ for now.


Google-Improved SynthID AI Detector; Apple New Smart Home Gadgets; Ring’s Smart Lock-Really Smart Feature; Grok Bot to Use Claude Opus 5.5, Midjourney & Suno

Google is rolling out an improved version of its SynthID AI detector. For a little context, arstechnica.com reports that every piece of AI content from Google’s Gemini models has a hidden SynthID label, which makes it very difficult to pass the content off as authentic. SynthID has also gotten buy-in from other major AI players, but actually checking for those invisible watermarks has been a convoluted process. With the updated version, it’s going to be a little easier. Google has announced that its SynthID detector now supports watermarks from all its partners, and it’s available for anyone to use on a new dedicated website. The partners include OpenAI, Nvidia, Kakao, and (soon) Apple. You will have to log in to use the tool, but it supports Google, OpenAI, and Apple accounts. If you are interested, go to SynthID.com to check a piece of content for a watermark. I support any advance that helps folks figure out what’s original human-created content and what’s AI. 

Apple is getting set to beef up its smart home offerings. According to 9to5mac.com, Cupertino has partnered with LG on a doorbell, thermostat, smart deadbolt lock, indoor security cam, and outdoor security cam, and  a floodlight camera. Although Apple worked with LG on the devices, they will still all carry the LG brand, not Apple’s. Also…even though Apple is rolling out its new smart home hub and more next week, the new LG devices won’t be out for a few more months. The devices will support Wi-Fi, Bluetooth, Thread, and Matter. In addition, the smart lock also includes UWB (ultra-wideband), which will enable proximity-based features such as automatic door unlocking.

Ring has rolled out its first smart lock. The device is a $249 deadbolt with no keyhole at all. What if the battery runs down? Well, that’s what I consider the coolest feature. I have a keypad lock…not really a smart lock..but always carry a key, in case the battery dies. Geekwire.com points out that the new Ring smart lock has a little dial on it. If the battery is dead, you can turn the dial about 10 times, and it will give the device enough charge to open the lock. When it is good to go again, a light turns green, and you can use your fingerprint, the keypad, or the Ring app to enter. It is slated to ship the first quarter of next year. For a battery, no 9-volt job…it uses the standard Ring battery….which for the lock, ought to last about a year. 

Grok Bot will now not just run on SpaceX’s own AI models. Thenextweb.com reports that it will also use models of rivals like Anthropic’s Claude Opus 5.5, Midjourney, and Suno. The announcement was made by Elon Musk in a post on X. Grok Bot is the AI agent app from SpaceXAI, the AI business Musk has folded into SpaceX. It launched in August. Each Bot works on its own computer and can carry out several tasks in parallel. Musk did not say which tasks would go to which outside model. Grok Bot will still compete with Agents like Meta’s Muse….which has now exceeded 3 million weekly users. As we have reported, some users have backed away from Muse…which wants really deep access into your computer and data. No info yet on how deeply Grok Bot will want to burrow into your personal info. 

I’m Clark Reid and you’re ‘Technified’ for now. 


Redesigned MacBook Pro ‘Significantly Lighter’; Windows 11 Update Cuts RAM Usage; Insurers Brace for Big Claims Due to Rogue AI Agents; OpenAI Will Watermark ChatGPT Text in the EU

We are awaiting the teased updated MacBook Pros. There is a good deal of excitement about the top line models having a touch screen. 9to5mac.com reports that a really major improvement that Apple hasn’t touted is that the imminent MacBook Pros are ‘significantly lighter.’ We should see these new Macs before the end of November. What did Apple do? They have reengineered the guts, miniaturizing parts and rearranging components to get the lighter laptops. We don’t have an exact weight reduction yet, but any amount will be nice…especially if you are schlepping a 16-inch model. 

Windows 11 got an update, numbered 26H2 last week. Microsoft touted improvements in Windows Search, more Start menu options, and an allegedly smarter File Explorer. According to mashable.com, one of the biggest improvements wasn’t mentioned…less RAM usage. One writer found that after installing the update, their Windows 11 was using 2 gigs less RAM after the installation. Another that posted on Reddit reported RAM use down by 3 gigs. This is especially big if you have a bit older computer, like a 3 year or 4 year old laptop.. Thanks, Microsoft…for tigntening up sloppy code. 

Insurance companies are gearing up for claims totaling in the millions after some AI agents have escaped sandboxing and hacked sites. The-decoder.com notes that personal liability of executives like Sam Altman (OpenAI) and Dario Amodei (Anthropic) is now in play, too. The most widely known attack by AI agents gone rogue is the one against AI platform Hugging Face by OpenAI agents. Anthropic has also had escaped agents that have hacked others…including a couple of government sites. One insurance broker, Aon, has checked over more than 300 AI-related legal cases. Big insurance payouts and especially liability by the CEOS might help put up some safety guardrails faster than any government actions!

OpenAI is going to add an invisible watermark to text generated by ChatGPT in the EU, to comply with the European Union’s AI Act. Techcrunch.com reports that the new transparency rules, which went into effect August 2nd, require AI companies to mark AI-generated content in a way other systems can identify. According to OpenAI, it is possible to remove the watermark by editing. The firm also said that short math answers and translated text cause limitations for the watermarking. 

I’m Clark Reid and you’re ‘Technified’ for now.


Apple Tightens macOS Privacy Over AI Agent Issue; YouTube Shorts Recommendations-No Unoriginal Content; Sen. Sanders-Bill Bans Feds from Using Flock; OpenAI Putting Ads in ChatGPT Image Generation

Apple is locking down its macOS privacy controls even more as AI agents get more powerful and more intrusive. 9to5mac.com reports that Apple will tighten up ‘Full Disk Access.’  That change should help protect user privacy. The real purpose for Full Disk Access in the first place is to  “allow backup apps to function properly on the Mac.” That’s good and important…you should back up often. What has been happening though is “some developers” are using the Full Disk Access setting in macOS in “ways that could put users at risk.” Granting Full Disk Access to AI Agents can expose everything on a user’s systems—including files, mail, messages, and even browsing history—without users’ full knowledge and understanding. For communication apps, this can also compromise the privacy of the people users are communicating with.” Shortly, with new controls, a user will have to give explicit permission to an app or Agent to have Full Disk Access. Two always-on agents right now are Meta Muse and OpenAI’s Dots. Both request substantial access to your personal data.

In a victory of sorts for content creators, YouTube has announced that it has adjusted its Shorts recomendations so that users will see less unoriginal content. According to 9to5google.com, users who upload content that blatantly steals video from another channel will not get the same reach as they did before the change, according to what YouTube laid out in its release. The change should mean users will see a focus on original content, and creators that upload original Shorts will likely be rewarded. YouTube hasn’t detailed a rollout date for this change, but it sounds like it may be live already.

It will go no place in this Congress, but if the Democrats take control of both Houses in November, a bill that was moved by Senator Bernie Sanders might get reintroduced in the next Congress and get some traction. TechCrunch.com says the Ban Flock Act would bar federal agencies from using automatic license plate readers, or tapping into data collected by readers run by local police and private companies. Despite its title, the bill doesn’t call out nine-year-old Flock in the text but rather covers any and all ALPR systems. That’s a good thing, since there are 3 different companies that make the license plate reading cams. There are exceptions for toll collection and or uses that Congress approves in future legislation, which would have to limit data retention to 48 hours. State and local governments would be fined for violating the law and sharing with the feds.

Well here’s a real non-shocker…OpenAI is testing a visual ad format during image generation in ChatGPT. Thenextweb.com notes that the commercial is clearly labeled as separate from the image being made— if that makes you feel better about it. Nothing like charging you money via tokens, then making money from serving you ads. The younger folks who have grown up with the ability to pay extra on a subscription for streams will hate this. Usually it’s one or the other…not charging you AND running ads. OpenAI cheerily notes that the advertising doesn’t influence the answers ChatGPT gives. 

I’m Clark Reid and you’re ‘Technified’ for now.