Apple reveals touchscreen MacBook Pro as Google ships Genie 4 and Gemini 4 Carbon

Apple unveils a touchscreen MacBook Pro with a brighter OLED display

John Coogan attended Apple's "Pro and Tell" event in person and reported that the headline reveal was a touchscreen MacBook Pro, alongside an OLED display upgrade. He praised Apple executive John Ternus's stage presence, saying "he is so good on stage" and describing the keynote pacing as tight, with only 15 seconds spent recapping the iPad Mini and the M6 MacBook Pro before moving to the touchscreen announcement. The hosts noted Apple did not display pricing on screen during the keynote, a departure from past events, which they attributed to rising component and memory costs. Pricing on apple.com showed the 14-inch model starting around $2,900 and the 16-inch around $3,400, with a leasing option available over 36 months.

The hosts also highlighted the return of the glowing Apple logo on the back of the laptop screen, which had been removed in past models reportedly to save costs. Coogan speculated this cost "like two bucks per unit" and said it had become proportionally cheaper to include now that overall laptop prices have risen. He noted audience applause for the glowing logo was "the biggest... of everything," even surpassing reaction to the touchscreen feature.

Google ships Genie 4 world model and Gemini 4 Carbon benchmark leader

Jordi Hays discussed Google's Genie 4, described by Sundar Pichai as "our most realistic and interactive world model," capable of steering dynamic environments in real time, generating native audio, and simulating photorealistic physics. It is currently in research preview, with Hays noting he does not yet have access and must apply for early access. The hosts floated potential use cases including content creation, virtual backdrops, and real-time scene compositing.

Separately, Google released Gemini 4 Carbon, which Pichai said "ranks number one on the recently updated V5 of the artificial analysis intelligence index at 59," scoring 88.2% on SEC Bench Pro and 54.8% on Exploit Gym for cybersecurity evaluations. The hosts joked about Google's atomic-number-based model naming scheme (Argon, Carbon) compared to OpenAI's celestial naming (Terra, Luna, Soul, Astra), with Coogan predicting, jokingly, that a future "helium model... could be the biggest model from Deep Mind in years."

OpenAI turns to Google's TPUs amid record compute leasing deals

The hosts cited a SemiAnalysis post from Dylan Patel noting OpenAI is reportedly discussing both GPU-as-a-service and TPU-as-a-service arrangements with neoclouds, including what Patel called "a 1 gigawatt TPU commitment." Patel's post also described an "8 gigawatt multiphase commitment in Ohio" for OpenAI, with Nvidia "backstopping 4.25 gigawatt," which he called "the largest single leasing commitment we have ever seen." Anthropic was reported to have leased 1.7 gigawatts this quarter, including 360 megawatts in Queensland, Australia, signed September 16, with one gigawatt backstopped by Google.

Coogan framed this as evidence against a "car-like" oligopoly market in AI chips, arguing instead that different chips suit different workloads: "rising tide will lift all boats." Separately, the hosts noted a listing on eBay for a 300-rack supercomputer powered by 9,600 Nvidia A100 GPUs, which they estimated originally cost around $300 million and is now priced near $60 million, though they judged "fair market value for this would be closer to 50 maybe 55 million."

Malaysia's compute buildout and its role as an AI trading hub

The hosts revisited a recurring topic on the show: Malaysia's outsized data center compute capacity relative to its size. Coogan recalled the framing from an earlier Wall Street Journal story describing Malaysia as "the Switzerland of AI," able to buy compute from the US and resell access in China. He speculated that Malaysian neoclouds racking Nvidia chips may be used by Chinese AI labs for training, given fewer trade restrictions, though he stopped short of confirming this as fact.

Reports of a $30 billion valuation for AI agent startup Instinct

A post from Lucas Grubber claimed Instinct is raising at a $30 billion valuation after moving through funding rounds from $100 million to $500 million to $2.5 billion to $10 billion in under a year, with Grubber arguing this signals readers should "pivot aggressively" into private markets. The hosts were skeptical of judging the claim, noting Instinct lacks a consumer app, unlike competitor Muse, which Hays said has "380,000 ratings, almost all five stars, 4.9 average" and ranks number one in the App Store's productivity category.

The hosts compared Instinct's trajectory to Cursor's funding history, noting Cursor had "north of a billion dollars of revenue" when it raised at a similar valuation, while Instinct's revenue and user metrics remain largely unknown.

"It's just really tough when you don't have the information... everyone else is giving takes based on how they feel about the number going up really fast."

— the host. They also discussed whether Instinct could eventually be acquired by a company like Apple or Amazon, similar to Cursor's acquisition, while questioning how such a deal would be treated under antitrust scrutiny given what one host called "the most competitive tech category we've had."