Overview
Google dominated the day with Gemini 4 Carbon, strong early benchmark rankings and an earnings report that fell short of forecasts. The wider conversation was about what faster AI development means for technical skills and businesses slow to adopt agents. Elsewhere, SpaceX completed engine tests for its next Starship flight, and eye-catching Apple prices prompted comparisons with used cars.
The big picture
Google launches Gemini 4 Carbon
Google introduced Gemini 4 Carbon just two weeks after Argon, claiming gains in cybersecurity, coding and tasks that require sustained work over many steps. The company says Carbon tops version 5.0 of the Artificial Analysis Intelligence Index.
Arena offered another view of its performance: Carbon’s High variant debuted first in Text Arena, second in Agent Arena and fourth in Code Arena: WebDev. That is a strong opening, though the different placements are a reminder that no single leaderboard tells the whole story.
Alphabet earnings miss forecasts
According to figures shared by unusual_whales, Alphabet reported earnings per share of $2.12 against an estimate of $3.01. Revenue came in at $127.5 billion, slightly below the $127.9 billion forecast.
The revenue gap was small; the earnings gap was much larger. The post does not explain what drove that difference, so it would be premature to pin it on AI spending or any particular part of the business.
Sundar Pichai says AI is speeding up Google’s model development
In earnings-call remarks shared by Andrew Curran, Sundar Pichai said AI is helping Google increase its training and launch speed. He expects further acceleration over the next twelve months, citing “RSI”.
The notable claim is that AI is already helping the process used to build and release new models. The quoted extract gives no measurement of that contribution, leaving an important question: how much faster can this development cycle become?
Thariq questions the anxiety around technical skills
Thariq, posting as trq212, described a curious divide: technical people worry that their skills are becoming useless, yet non-technical people wish they knew more so they could use agents better.
Both reactions can exist at once. Automating parts of a task does not remove the need to define the problem, check the result or work out what went wrong. Technical knowledge may become less about writing each line and more about knowing whether the work holds up.
Mark Cuban warns companies about slow AI adoption
In a clip shared by a16z, Mark Cuban argued that weak AI adoption leaves established organisations exposed to ambitious newcomers. His phrase for the potential winner was the “trillionaire in the basement”.
That is a prediction, not a business plan. The more practical point is his insistence that chief executives understand how AI affects their own companies, rather than treating it solely as an IT purchase.
SpaceX completes Ship 42 engine tests
SpaceX says Ship 42 completed a full-duration static fire using all six engines ahead of Starship Flight 15. It also fired a single Raptor engine to simulate a deorbit burn.
These are ground tests, not proof of flight readiness on their own. Still, they test two important parts of the planned mission before the vehicle leaves the launch site.
Ivan Fioravanti reports a €27,000 Apple configuration
Ivan Fioravanti reported an Apple Store listing in Italy for an M5 Ultra configuration with 512GB of memory and a 4TB SSD, priced at €27,000. His post lists availability from 23 November.
At that price, the useful question is not whether it makes sense as an ordinary desktop purchase. It is whether a specific professional workload can justify the memory capacity and cost. The listing details here come from Fioravanti’s report rather than a separate Apple announcement.
Sawyer Merritt finds a used Tesla priced below a MacBook Pro
Sawyer Merritt compared a $14,190 listing for a 2020 Model 3 Standard Range with a fully configured 16-inch OLED MacBook Pro priced at $14,549. The car had 95,000 miles and, according to the post, a clean title with no accident or damage history.
It is a memorable price comparison, not a like-for-like buying decision. A used car brings running costs and wear; a top-spec laptop carries its own premium. What the example captures is how far apart the depreciation of a vehicle and the pricing of high-end computing can look.
