Tech Brew Ride Home - GPT-5.2 As OpenAI’s Attempt To Change The Narrative

Episode Date: December 12, 2025

Ok, fine, says Sam, here’s a new GPT model so you’ll hopefully stop saying we’re behind. Broadcom as another AI bellwether. Now that Disney is in bed with OpenAI, they’re ceasing and desisting Google. And, of course, The Weekend Longreads Suggestions. OpenAI Launches GPT-5.2 as It Navigates ‘Code Red’ (Wired) GPT-5.2 is OpenAI’s latest move in the agentic AI battle (The Verge) Trump threatens funding for states over AI regulations (Reuters) Broadcom beats on earnings and revenue, says AI chip sales will double in current quarter (CNBC) Disney Accuses Google of Using AI to Engage in Copyright Infringement on ‘Massive Scale’ (Variety) Weekend Longreads Suggestions: Want This Hearing Aid? Well, Who Do You Know? (Wired) Tech bros head to etiquette camp as Silicon Valley levels up its style (The Washington Post) Why AGI Will Not Happen (Tim Dettmer) 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 Techrew right home for Friday, December 12th, 2025. I'm Brian McCullough today. Okay, fine, says Sam. Here's a new GPT model, so you'll hopefully stop saying we're behind. Broadcom as another AI bellwether. Now that Disney is in bed with OpenAI, they're ceasing and desisting Google, and of course, the weekend long-read suggestions. Here's what you missed today in the world of tech. You may have noticed that your customers love, webinar, and video content, but you also might have noticed that if you've ever put a webinar or video together, then, it can eat up a lot of your time and budget. But now, thankfully, there's a singular tool that can streamline your team's video and webinar workflows, Wistia.

Wistia can scale your content output with AI-powered tools that help you create, edit, and repurpose videos and webinars fast. And speaking of webinars, you can host engaging, easy-to-set-up webinars in Wistia, complete with built-in analytics. With Wistia, you don't have to pay for multiple video tools, hop between platforms, or constantly re-upload files. Create, edit, collaborate, and publish all in one place. Head to wistia.com slash brew to learn more. That's WISTA.com slash brew. With Wistia, you can expect less work and more plays. Well, I told you I expected OpenAI to do something

before the end of the year to try to arrest their sort of narrative slide. They sort of had to, so, quoting Wired. OpenAI has introduced GPD 5.2, its smartest artificial intelligence intelligence model yet, with performance gains across writing, coding, and reasoning benchmarks. The launch comes just days after CEO Sam Altman internally declared a code read, a company-wide push to improve chat GPT amid intense competition from rivals. We announced this code read to really signal to the company that we want to marshal resources in one particular area, and that's a way to really define priorities, said OpenAI's CEO of Applications, Fiji Simo in a briefing with reporters on Thursday, we have had an increase in resources focused on chat GPT in general.

Much like the company's recent model launches, GPT 5.2 is shipping as a series of models, instant, which responds faster and is better for information finding, thinking, which excels at coding, math, and planning, and pro the most powerful tier of OpenAI's models that delivers higher accuracy on difficult questions. OpenAI calls GPT 5.2 its best model yet for everyday professional use. GPD 5.2 thinking notched the highest scores to date on GDPVOL, an OpenAI benchmark that compares performance between AI models and human professionals across 44 real-world occupations. The company says the model beat human professionals in over 70% of tasks and completed them 11 times faster. OpenAI's post-training lead, Max Schwerzer, says the new release should also

offer a substantial reduction in hallucinations. The company says GPD 5.2 thinking hallucinated 30, 38% less than GPD 5.1 on benchmarks measuring answers to factual questions. The company is bringing GPT 5.2 to both chat GPT users and developers on OpenAI's API product. OpenAI says the new series of models, quote, brings clear gains across every day and advanced use cases, end quote. And quoting the verge. OpenAI says the GPD 5.2 model series, which includes the instant thinking and pro models, is better at, quote, creating spreadsheets, building presentations, writing code,

perceiving images, understanding long contexts, using tools and handling complex multi-step projects. Aidan Clark, a VP of Research at OpenAI, said in the briefing that the team gave a senior immunology researcher access to GPT 5.2 Pro. He said that when the researcher asked the model to generate the most important unanswered questions about the immune system, they said it produced, quote, sharper questions and stronger explanations for why those questions matter than any other frontier model. In a blog post, OpenAI also said GPT 5.2 is better for AI agents' workflows, part of the ever-intensifying battle between AI companies to offer the most efficient and useful AI agents. The vision is for ChatGPT to be the best possible personalized assistant. Max

Schwitzer, a researcher at OpenAI, said during the briefing. The company also said its thinking model hallucinates less than its predecessor, making sure to highlight that the distinction makes it more useful for professionals looking for trustworthy agentic AI tools. Notion, Box, Shopify, Harvey, Zoom, and Databricks were among the pre-release testers, and they received access a couple of weeks ago, per OpenAI, end quote. Simon Willison says the GPD 5.2 models match GPT5 and 51 with a 4,000 context window and 128,000 max output tokens, but have a newer knowledge cutoff of August 31st, 2025 versus September. September 30th, 2024 previously.

Separately, quoting the verge, Fiji Simo, OpenAI, CEO of applications told reporters during a Thursday briefing about GPT 5.2 that she expects adult mode to debut within Chad GPT in the first quarter of 26, adding that the company wants to get better at age prediction before introducing the new feature. OpenAI is currently in the early stages of testing its age prediction model, which is designed to automatically figure out when to apply certain safeguards and content restrictions. for users under 18. Seymot said during the briefing that the company is already testing the model in certain countries to gauge its ability to identify teens and not misidentify adults, something that the

company wants to get right before it debuts the adult features, end quote. President Trump has signed an executive order aimed at preempting a growing number of state AI laws, quoting Reuters, we want to have one central source of approval. Trump told reporters flanked by top advisors, including Treasury Secretary Scott Bessett, arguing that 50 different regulatory regimes hamper the growth of the nascent industry. To win, United States AI companies must be free to innovate without cumbersome regulation, the order said, adding that the current patchwork of different regulatory regimes makes compliance more challenging, especially for startups. The order will give the Trump administration

tools to push back on the most, quote, onerous state regulations, said White House AI advisor David Sachs. The administration will not oppose rules governing AI that relate to child safety, he added. It directs the Secretary of Commerce to evaluate state laws for conflicts with Trump's AI priorities and to block those states in conflict from accessing the $42 billion broadband equity access and deployment fund. Democratic Representative Don Bayer, who co-chairs a bipartisan caucus on AI, said the order would squelch safety reforms passed by states and create a, quote, lawless Wild West environment for AI companies that puts Americans at risk. He warned that the order would reduce the likelihood of congressional action and likely violated the 10th Amendment,

which says that any powers not specifically given to the federal government belong to the states or the people. Trump's order called for his administration to work with Congress to craft a national standard that forbids state laws which conflict with federal policy, protects children, prevents censorship, respects copyrights and protects communities. Until such a standard was in place, the order called for actions to, quote, check the most onerous and excessive laws emerging from the states that threatened to stymie innovation, end quote. On the, are we in an AI bubble slash is OpenAI in Troublefront? Broadcom reported Q4 revenue up 28% year on year to $18.02 billion and forecast Q1 revenue above estimates with AI chip sales doubling to $8.2 billion.

Quoting CNBC, along with Nvidia Broadcom has been the other major winner among U.S. semiconductor companies from the AI boom. The company's stock prices up 75% so far. far in 2025 after doubling last year as its custom chips, such as Google's tensor processing units are gaining increasing traction in the market as a rival to Nvidia's graphics processing units. Investors are closely watching to see Broadcom confirm that those companies are continuing to engage and are on track to buy and deploy custom chips. Broadcom revealed during a September earnings call that it had signed a customer that had placed a $10 billion order for custom chips. At the time, Broadcom didn't say who it was, but on Thursday, CEO Haq Tan, revealed that the mystery customer was AI Lab Anthropic,

which placed an order for the latest Google Tensor Processing units. We received a $10 billion order to sell the latest TPU ironwood racks to Anthropics, said Tan, speaking on Broadcom's fourth quarter earnings call on Thursday. He also said Anthropic had placed an additional $11 billion order with Broadcom in the company's latest quarter. While Broadcom typically doesn't disclose its large customers, TAN's September remark drew significant investor attention amid the AI infrastructure boom. A Broadcom official told CNBC in October that the mystery customer wasn't open AI, which has its own agreement to purchase chips from the chipmaker. Broadcom makes custom chips called A6, which some experts believe are more efficient for certain artificial intelligence algorithms than the market-dominating chips from Nvidia. Broadcom helps make Google's TPUs, and last month, the search company,

brag that it trained its state-of-the-art Gemini 3 model entirely on its TPUs. The chipmaker calls its custom AI chips XPUs, and on Thursday, Tan said his company was delivering entire server racks, not just chips, too Anthropic, which is Broadcom's fourth XPU customer. Broadcom on Thursday also said that it had secured a fifth customer for its custom chip business. That customer placed a $1 billion order during the fourth quarter, but once again, Broadcom did not reveal the customer. It's a real customer, and it will grow, Tan said, 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. slash college PC. 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. There's one big addendum to yesterday's news of that Disney deal with Open AI, quoting variety. Now that Disney has gone into business with Open AI, the mouse house is accusing Google of copyright infringement on a massive scale using AI models and services to quote commercially exploit and distribute infringing images

and videos. On Wednesday evening, attorneys for Disney sent a cease and desist letter to Google demanding that Google stop the alleged infringement in its AI systems. Google is infringing Disney's copyrights on a massive scale by copying a large corpus of Disney's copyrighted works without authorization to train and develop generative artificial intelligence models and services and by using AI models and services to commercially exploit and distribute copies of its protected works to consumers in violation of Disney's copyrights reads the letter to Google's General Counsel from law firm, Jenner, and Block on behalf of Disney. According to the letter, which Variety has reviewed, Disney alleges that Google's AI systems and services infringe

Disney characters, including those from Frozen, The Lion King, Moana, The Little Mermaid, Deadpool, Guardians of the Galaxy, Toy Story, Brave, Ratatooie, Monsters, Inc., Lilo and Stitch, Inside Out, and franchises such as Star Wars, The Simpsons, and Marvel's Avengers and Spider-Man. In its letter, Disney included examples of images it claims were generated by text prompts in Google's AI apps, including of Darth Vader. The allegations against Google follow cease-and-assist letters that Disney sent earlier to meta and character AI, as well as litigation Disney filed together with NBC Universal and Warner Brothers discovery against AI company's Mid-Jurney and Minimax, alleging copyright infringement. As for comment, a Google spokesperson said, we have a longstanding and mutually beneficial relationship with Disney and will continue to engage with

them. More generally, we use public data from the open web to build our AI and have built additional innovative copyright controls like Google extended and content ID for YouTube, which gives sites and copyright holders control over their content. According to Disney, the company has been raising its concerns with Google for months, but says Google hasn't done anything in response, and that, if anything, Google's infringement has only increased during that time. Iger, Disney's CEO in an interview with CNBC Thursday, said, well, we've been aggressive at protecting our IP, and we've gone after other companies that have not honored our IP, not respected, our IP, not valued it. And this is another example of us doing just that. Iger said Disney had been in

discussions with Google basically expressing our concerns about its AI systems alleged infringement, and ultimately because we didn't really make any progress, the conversations didn't bear fruit. We felt we had no choice but to send them a C-Sy synthesis letter. Disney's letter to Google demands that Google immediately cease further copying, publicly displaying, publicly performing, distributing, and creating derivative works of Disney's copyrighted characters in outputs of Google's AI services, including through YouTube's mobile app, YouTube shorts, and YouTube, end quote. Time for the weekend long read suggestions. First up from Wired, a profile of AI hearing aid startup, Fortell, which has raised $150 million so far and is targeting affluent clients in New York

City with a $6,800 device that uses a custom chip. Quote, a secret is percolating at dinner party, salons, and cocktail gatherings among the August New York City elite. It's whispered in the circles of financial masters of the universe, Hollywood stars, and owners of sports teams. Have you heard about foretell? Many haven't, or if they did hear, they might not have made out the words through noisy cross-conversations. Once they do know, particularly if they're boomers, they want it desperate. foretell is a hearing aid, one that claims to use AI to provide a dramatically superior oral experience. The chosen few included in its beta test claim that it seems to top the performance of high-end devices they'd been unhappily using.

These testers have made pilgrimages to Fortel's headquarters on the fifth floor of a we-work facility in New York City's trendy Soho neighborhood, where they were fitted for the hearing aids, which from the outside look pretty much like standard over-the-ear, tear-drop-shaped devices, but the big moment comes when a foretell staffer takes them down to the street level. There among street clatter, honking cabs and delivery trucks backing up to luxury stores, they are asked to conduct a conversation with a foretel worker. Two other employees stand behind them, adding their own loud discourse to the urban cacophony. Despite the din, the testers clearly make out what the person in front of them is saying. The clouds lift, angels croon. This was so incredible that I burst into tears, says Ashley Tudor.

one of the seemingly few beta testers who isn't famous or powerful, though she is married to a venture capitalist, end quote. Then from the Washington Post, a look at a so-called etiquette camp for young founders in San Francisco, which teaches them how to dress, act, and talk like, well, the modern founder. Quote, tech founders have traditionally eschewed business norms as they worked to upend the status quo, sticking with hoodies and jeans, even as dorm room dreams became billion-dollar corporations, and donning suits only when hauled before a court or Congress. Scruffy grooming and unrefined social skills became hallmarks of the gifted entrepreneur laser-focused on changing the world.

But the expectations of how a tech founder looks and acts and the role models of Silicon Valley's success are shifting as the industry gains power and influence. Tech employers and investors want the full package, said Caroline Samard, dean of Northeastern University's Silicon Valley campus. That means people who have technical chops and who can collaborate their good communicators, and they have critical thinking, she said. Rising tech leaders are becoming more concerned with how they present themselves. Victoria Hitchcock, a Bay Area-style consultant who has helped tech clients who visited the White House this year,

said people in the industry now request major changes to their entire persona. In her 15 years working with tech clients, she has often had to point male founders towards simple areas to improve like personal grooming, but they will now proactively ask how to get rid of under-eye circles or for the best hair replacement program, Hitchcock said, end quote. And finally, from my, I don't have a personal opinion on this, but I do try to present you every angle of a debate file. From Tim Detmer's, a piece that makes its point right in its title, Why AGI will not happen.

Detmers basically argues that forecasts of imminent AGI or superintelligence ignore the physical reality of computation. The main bottleneck is moving information, memory, bandwidth, latency, so shrinking transistors makes arithmetic cheaper while making memory relatively more constraining. Transformer-style models are already close to physically efficient architectures. Because improvements here face diminishing returns, he claims linear capability gains require exponentially more resources. That mattered less when GPU efficiency rose quickly, but he argues GPUs have large plateaued since around 2018, so further scaling will become prohibitively expensive and short-lived. In that world, winner-take-all-frontier bets look fragile, and value shifts to broad deployment

and adoption. He also says true AGI would need economically meaningful physical work, yet robotics data is costly and most useful robots will remain specialized. So runaway recursive self-improvement is unlikely. The future is incremental progress and widespread economic diffusion of good enough AI into practical applications, not a sudden AGI breakthrough. So I think there will be a bonus episode for you this weekend. Like I did last year, I went on the Newsworthy podcast to count down what I thought were the top 10 tech stories of the year.

If they get me the audio by tomorrow, I'll post it for you as a bonus episode. If they don't, I won't. Talk to you on Monday. All. Pay off your home, travel for life, drive a Ferrari. In celebration of the world premiere of the Monopoly Big Board Buckslot machine by Aristocrat Gaming, Yamava Resort and Casino at San Manuel is giving one person a $1.6 million dream package.

The biggest prize in Yamava's history. Club Serrano members can earn daily instant prizes and secure a spot in the finale May 29th. Don't pass go and own it all. Only at Yamava, celebrating its 40th anniversary. You win? Details at Yamava.com must be 21-20. Please gamble responsibly.

Monopoly is a trademark of Hasbro. Hasbro is not a sponsor of this promotion. Thank you.