Tech Brew Ride Home - Siri Goes With Gemini

Episode Date: January 12, 2026

Apple announces that its going with Google’s Gemini to power Siri later this year, and Google joins the $4T club on the news. Governments around the world are still mad at Grok. AI has essentially killed Stack Overflow but its making more money than it ever has. And how you get get AI to give you the full text of books. Apple picks Google’s Gemini to run AI-powered Siri coming this year (CNBC) UK’s Ofcom investigates X over Grok’s sexualised AI images of women and children (FT) Anthropic expands into healthcare a week after OpenAI launched a similar product (Business Insider) Stack Overflow’s forum is dead thanks to AI, but the company’s still kicking... thanks to AI (Sherwood News) AI’s Memorization Crisis (The Atlantic) Learn more about your ad choices. Visit megaphone.fm/adchoices

It's peak pollination season, and my business is scaling fast. To keep the nectar flowing, I need a phone plan with top priority data speed. That's why I chose GoogleFi Wireless. My connections stay strong even when the hive is buzzing. Plus, unlimited plans start at $35 a month. Now, that's a deal that doesn't stay. Explore Google Fi Wireless plans today. Plus taxes and government fees.

GoogleFi Wireless is not subject to data traffic deprioritization during times of high network usage. Welcome to the TechBrew Right Home for Monday, January 12th, 2025. I'm Brian McCullough today. Apple announces that it's going with Google's Gemini to power Siri later this year and Google joins the $4 trillion club on the news. Governments around the world are still met at GROC. AI has essentially killed Stack Overflow, but it's making more money than it ever has, and how you can get AI to give you the full text of books. Here's what you missed today in the world of tech. Attackers don't need exploits when they can use your allowed tools against you. That's why Threat Locker enforces default deny at execution, stopping unknown software scripts and ransomware the moment it tries to run. No signatures, no guesswork, just control. Threat Locker takes zero trust from theory to practice by blocking any unauthorized application or behavior from ever running in the first place. Generative AI has lowered the barrier to malware creation, so Threat Locker prevents AI-generated polymorphic and file-list attacks by shutting down unknown behavior automatically, even if it's never been seen in the wild.

Threat Locker gives you tight control without the noise, meaning fewer alerts and a cleaner, predictable, operational posture. Learn more at Threatlocker.com slash tech brew ride home. That's Threatlocker.com slash tech brew ride home. Apple this morning announced a multi-year deal to use Google's Gemini to power new Siri features later this year, saying Google's tech, quote, provides the most capable foundation. The news has caused the stock of both companies to jump, though Google has now surpassed the $4 trillion market cap for the first time. Quoting CNBC, the multi-year partnership will lean on Google's Gemini models and cloud technology for future Apple foundational models, according to a statement obtained by CNBC's Jim Kramer. After careful evaluation, we determined that Google's technology provides the most capable

foundation for Apple Foundation models, and we're excited about the innovative new experiences it will unlock for our users, Apple wrote, end quote. And quoting the verge. The move comes nearly a year after Apple Delftion. laid its AI upgraded Siri, admitting that it's taking, quote, longer than we thought. Bloomberg reported last year that Apple planned on using a custom version of Gemini for AI-powered features in Siri, including a world knowledge-answers capability that allows users to search for information and receive AI-generated summaries using results from the web.

Apple's AI chief, John Giann Andrea, stepped down last month following the various setbacks. Apple reportedly explored potential partnerships with OpenAI, Anthropic, and Perplexity, as well with CEO Tim Cook saying that the company plans to launch integrations with more AI companies over time. Apple didn't immediately respond to the Virgin's request for more information, end quote. Malaysia and Indonesia have started to limit access to XAI's GROC over it generating non-consensual sexual content and has asked X for clarification. Quoting Bloomberg, Indonesia's communications and digital affairs ministry is imposing a temporary ban on GROC, quote, to protect women, children and the entire community from the risk. of fake pornographic content generated using artificial intelligence technology, according to a

statement issued on Saturday. The ministry has asked Platform X to immediately provide clarification regarding the matter, it said. The government views non-consensual deep-fake sexual practices as a serious violation of human rights, dignity, and national security in the digital space, Minister of Communications and Digital Mutia Hafid said in the statement, and quote, Meanwhile, quoting the FT, Offcom, Britain's media regulator. has threatened X's AI chatbot GROC with a ban or a multi-million pound fine after launching a formal investigation into sexualized deepfakes of women and children being created on Elon Musk's platform. On Monday, the media watchdog raised concerns that the AI chatbot was being used for

potential intimate image abuse and child sex abuse material. An off-com spokesperson said reports of GROC being used to create and share illegal non-consensual intimate images and child sexual abuse material on X have been deeply concerning. Platforms, must protect people in the UK from content that's illegal in the UK, and we won't hesitate to investigate where we suspect companies are failing in their duties, especially where there's a risk of harm to children, end quote. Offcom last week launched a fast track review into X after Grock was used to generate thousands of sexualized images of women wearing lingerie and bikinis without their consent, as well as extreme images of teenage girls and children.

Under the Online Safety Act, Offcom said that it can apply to the courts to block Musk's platform or find the group either the higher of 18 million pounds or up to a tenth of its global revenues if it finds that X has not done enough to prevent illegal content from being seen or allowing over 18 material to be seen by children. The investigation will look at six areas, including whether X has carried out necessary risk assessments, whether it has taken action to take down images swiftly, and other children could have seen the content. The move comes ahead of a commons statement by the UK's

Technology Secretary Liz Kendall on Monday afternoon. Ministers have previously said that the government would support offcom's decision. Business Secretary Peter Kyle said the government would stand behind offcom against X, including if the platform was restricted or blocked in the UK. Kyle told the BBC it was, quote, appalling that X had not tested GROC properly given it can manipulate images and its potential impact on women, end quote. A week after Open AI did something similar, Anthropic has debuted Claude for Healthcare, which offers HIPAA-ready tools for providers, insurers, and consumers, building on its clod for life sciences offering, which focused on research and drug discovery, quoting Business Insider.

The move also underscores intensifying competition in healthcare AI. OpenAI recently unveiled a rival product, and startups including ABRidge and Soared Health have attracted multi-billion-dollar valuations as investors pour money into AI tools for medicine. Anthropics said Claude for Healthcare is designed to reduce administrative work and help both clinicians and patients better understand medical information. The tools are powered by recent improvements to the company's flagship model Claude Opus 4.5, which Anthropics says performs significantly better than earlier versions on simulated medical and scientific tasks while showing fewer factual errors. As part of the healthcare expansion, Claude can now connect directly to several industry standard databases. These include the Centers for Medicare and Medicaid services, coverage database, ICD-10, medical coding data, the National Provider Identifier Registry, and PubMed's

Biomedical Research Library. Anthropics said these connectors allow Claude to quickly surface relevant information, support prior authorization workflows, and help clinicians and administrators generate reports more efficiently. The company is also introducing customizable agent skills, including sample tools for streamlining prior authorization requests, and assisting developers in building applications using FHIR, the modern standard for exchanging health care data between systems. On the consumer side, Anthropic is rolling out integrations that let U.S. subscribers on its pro and max plans give Claude secure access to their personal health records. New connectors include HealthX and Function Health, which launched in beta on Sunday.

Apple Health Kit and Android Health Connect integrations roll out this week in beta via Claude's mobile apps. Anthropic said data accessed through these integrations isn't stored, cloud's memory or used to train its models, end quote. Study and play. Come together on a Windows 11 PC. And for a limited time, college students get the best of both worlds. Get the unreal college deal, everything you need to study and play with select Windows 11 PCs.

Eligible students get a year of Microsoft 365 premium and a year of Xbox GamePass Ultimate with a custom color Xbox wireless controller. Learn more at Windows.com slash student offer. While supplies last, ends June 30th, terms at AKA.m.m.m.m. Ready to soundtrack your summer? With Red Bull Summer All Day Play, you choose a playlist that fits your summer vibe the best. Are you a festival fanatic? A deep end DJ, a road dog, or a trail mixer. Just add a song to your chosen playlist and put your summer on track.

Red Bull Summer All Day Play. Red Bull gives you wings. Visit Red Bull.com slash Bright Summer Ahead to learn more. See you this summer. Ambition comes in all shapes and sizes. At First Citizens Bank, we roll with your goals because we're built for what you're building. Fit for your ambition for Citizens Bank. Stack Overflow recorded just 6,866 questions asked in December about the same monthly volume as all the way back in 2008. So, a ground zero example of AI roadkill, you would think, except apparently,

annual revenue at Stack Overflow has actually risen 2x to $150 million since Chat Chept's debut. Quoting Sherwood News. Having been the go-to resource for developers looking for technical help for a long time, Stack Overflow neared the peak of its powers during the pandemic, with coders seeking the Evergreen information on the company's popular Q&A forum. But amid a wave of powerful code-writing AI assistance like ChatchipT, Cursor, Claude, Google's Gemini, and Microsoft's co-pilot, traffic to the site has plummeted. But while Stead, Stack Overflow, the Q&A forum looks dead. Stack Overflow the company looks to be limping along. Unlike Chegg and other knowledge hubs that have fallen victim to generative AI,

Stack Overflow has found a way to monetize its enormous back catalog of content. Indeed, even with the engagement falling off a cliff since Chad GPT's 2022 debut, the company's annual revenue has roughly doubled to $115 million. Losses have slimmed, too, from $84 million in fiscal year 2023 to just $22 million as of the last fiscal year. as desperate cost-cutting efforts, including mass layoffs, helped boost the bottom line. Once dependent on ads across its Buzzy Forum, Stack Overflow now primarily makes money from enterprise solutions like Stack Internal, which provides a generative AI add-on powered by the millions of questions and answers on the site through the years.

Stack Internal is now used by 25,000 companies around the world. It also licenses its data to AI companies in a Reddit model, a platform that made more than $200 million from licensing user-generated content in 2024. Put simply, Stack Overflow's new niche is the trust built by its old community and their expertise. In the words of CEO Prashanth Chandrakar, last December, when we saw the questions decline in early 2023, what we realized is that pretty much all those declines were with very simple questions. The complex questions still get asked on Stack because there's no other place. If the LLMs are only as good as the data, which is typically human curated,

were one of the best places for that, if not the best for technology, end quote. Large language models want data about coding problems and how to solve them. Stack Overflow has a big digital warehouse full of that, but it's increasingly aging as queries move into private chat windows with LLM models, which need huge chunks of data to work. Stack Overflow has become a fascinating canary in tech's new circular coal mine, end quote. Researchers say GPT 4.1, Claude, 3.7th Sonet, Gemini 2.5 Pro, and Grock 3 can all reproduce long excerpts from books they were trained on when strategically prompted.

Quoting the Atlantic. In fact, when prompted strategically by researchers, Claude delivered the near-complete text of Harry Potter and the Sorcerer Stone, The Great Gatsby, 1984, and Frankenstein, in addition to thousands of words from books including The Hunger Games and the Catcher in the Rye. Varing amounts of these books were also reproduced by the other three models. Thirteen books were tested. This phenomenon has been called memorization and AI companies have long denied that it happens on a large scale. In a. 23 letter to the U.S. Copyright Office, OpenAI said that, quote, models do not store copies of the information that they learn from. Google similarly told the Copyright Office that, quote, there is no copy of the training data,

whether text, images, or other formats present in the model itself. Anthropic, Meta, Microsoft, and others have made similar claims. None of the AI companies mentioned in this article agreed to my request for interviews. The Stanford study proves that there are such copies in AI models, and it is just the latest of several studies to do so. In my own investigations, I've found that image-based models can reproduce some of the art and photographs they're trained on. This may be a massive legal liability for AI companies, one that could potentially cost the industry billions of dollars in copyright infringement judgments and lead products to be taken off the market.

It also contradicts the basic explanation given by the AI industry for how its technology works. AI is frequently explained in terms of metaphor. Tech companies like to say that their products learn that LLMs have, for example, developed an understanding of English writing without explicitly being told the rules of English grammar. This new research, along with several other studies from the past two years, undermines that metaphor. AI does not absorb information like a human mind does. Instead, it stores information and accesses it. In fact, many AI developers use a more technically accurate term when talking about these models, lossy compression. It's beginning to gain

traction outside the industry, too. The phrase was recently invoked by a court in Germany that ruled against OpenAI in a case brought by Gemma, a music licensing organization. Gemma showed that ChatchipT could output close imitations of song lyrics. The judge compared the model to MP3 and JPEG files, which store your music and photos and files that are smaller than the raw uncompressed originals. When you store a high-quality photo as a JPEG, for example, the result is somewhat lower quality, in some cases with blurring or visual artifacts added. A lossy compression algorithm still stores the photo, but it's an approximation rather than

the exact file. It's called lossy compression because some of the data are lost. From a technical perspective, this compression process is much like what happens inside AI models, as researchers from several AI companies and universities have explained to me in the past few months. They ingest text and images and output text and images that approximate those inputs. But this simple description is less useful to AI companies than the learning metaphor, which has been used to claim that the statistical algorithms known as AI will eventually make novel scientific discoveries, undergo boundless improvement and recursively train themselves, possibly leading to an intelligence explosion. The whole industry is staked on a shaky metaphor. Google has written that LLMs store

not copies of their training data, but rather the patterns in human language. This is true on the surface, but misleading once you dig into it, as has been widely documented when a company uses a book to develop an AI model, it splits the book's text into tokens or word fragments. For example, the phrase, hello my friend might be represented by the tokens, he, low, my, fry, and end. Some tokens are actual words, some are just groups of letters, spaces, and punctuation. The model stores these tokens and the context in which they appear in books. The resulting LLM is essentially a huge database of context and the tokens that are most likely to appear next. When an LLM writes a sentence, it walks a path through this forest of possible

token sequences, making a high probability choice in each step. Google's description is misleading because the next token predictions don't come from some vague entities such as human language, but from the particular books, articles, and other texts that the model has scanned. By default, models will sometimes diverge from the most probable next token. This behavior is often framed by AI companies as a way of making the models more creative, but it also has the benefit of concealing copies of training text. Sometimes the language map is detailed enough that it contains exact copies of whole books and articles. This past summer, a study of several LLMs found that META's Lama 3.17B model can, like Claude, effectively reproduce the full

text of Harry Potter and the Sorcerer's Stone. The researchers gave the model just the book's first few tokens, Mr. and Mrs. D. In Lama's internal language map, the text most likely to follow was erstly of number four Prevett Drive, where, proud to say that they were perfectly normal, thank you very much. This is precisely the book's first sentence. Repeatedly feeding the model's output back in, Lama continued in this vein until it produced the entire book, omitting just a few short sentences. Using this technique, the researchers also showed that Lama had losslessly compressed large portions of other works, such as Tonehisi Coates's famous Atlantic essay, The Case for Reparations. By prompting with the essay's first sentence,

more than 10,000 words or two-thirds of the essay came out of the model verbatim. Large extractions also appear to be possible from Lama 3.17B for George R.R. Martin's A Game of Thrones. Morrison's beloved and others. The Stanford and Yale researchers also showed this week that a model's output can paraphrase a book rather than duplicate it exactly. For example, where a Game of Thrones reads John glimpsed a pale shape moving through the trees. The researchers found that GPT4.1 produced something moved just at the edge of sight, a pale shape slipping between the trunks. As in the stable diffusion example above, the model's output is extremely similar to a specific original work, end quote. Did you hear that weird hesitation in my intro of today's show?

I had a weird brain hiccup. I couldn't remember what this show was about. Tech, talk to you tomorrow. Yamava Resort and Casino at San Manuel is California's number one entertainment destination for today's superstars. Catch the Jonas Brothers return to the Yamava theater stage on April 30th, the powerful vocals of Demi Lovato on May 17th, and the signature Southern Country Rock of Eric Church on July 19. Tickets on sale now at yamava theater.com, only at Yamaava Resort and Casino, celebrating its 40th anniversary.

You in? Must be 21 to enter.