TBPN

Frontier AI models scored near-zero on ARC-AGI V3—GPT-5.5 at 0.43% and GPT-4.7 at 0.18%—indicating that despite rapid AI progress, a meaningful capability ceiling still exists on novel visual-reasoning tasks, which the hosts see as a reassuring counterweight to hype.

Amazon's Q1 2026 total revenue was $181.5 billion, up 17% year-over-year, with its advertising business generating $17.2 billion and its custom chips business crossing a $20 billion annualized run rate.

Outside the MAG-4, several AI-adjacent public companies trade at bubble-like multiples—AMD at 130x, Palantir at 220x, and at least one name at 900x earnings—suggesting valuation concentration risk is real but localized, not systemic across the whole mega-cap cohort.

The MAG-4 companies that reported earnings trade at 17–25x forward earnings—Meta ~17x, Google 17x, Amazon 24x, Microsoft 25x—compared to dot-com peak multiples of 73x for Microsoft, 200x+ for Cisco, and the NASDAQ as a whole trading at 200x, making the current AI leaders look reasonably valued by historical comparison.

Meta's daily active people (DAP) declined sequentially for the first time since the company began tracking the metric, though management attributed the drop to internet disruptions and WhatsApp restrictions in specific regions rather than structural user loss.

Meta reported Q1 2026 revenue of $56.3 billion, up 33% year-over-year, but raised its capex range by $10 billion on each end, now guiding $125–$145 billion, raising questions about whether incremental spend buys incremental compute or just offsets rising input costs.

AWS grew 28% year-over-year in Q1 2026, beating the consensus expectation of 25%, a meaningful re-acceleration signal at AWS's scale.

OpenAI's exclusivity with Azure has ended; OpenAI models are now available on AWS as well, which weakens Azure's competitive sales advantage but benefits Microsoft's equity stake in OpenAI as the model reaches more customers.

Microsoft's Copilot seat count grew from 5 million to 20 million, but remains a small fraction of the company's 450 million total paid Microsoft seats, indicating significant upsell runway.

Google Search revenue grew 19% year-over-year, effectively disproving the thesis that AI chatbots are cannibalizing core search revenue.

Google Cloud's backlog nearly doubled to over $460 billion, with more than half expected to be recognized as revenue within the next 24 months, signaling durable near-term growth.

Google Cloud revenue hit $20 billion in Q1 2026, up 63% year-over-year, massively beating analyst expectations of roughly 40% growth.

Regulatory fines in tech legal cases remain trivially small relative to company revenues, and Isaac argues this pattern means even landmark AI liability rulings are unlikely to be financially deterrent for large incumbents.

Trial testimony included significant negative commentary by Musk about Larry Page — specifically that he no longer trusts him — surfacing a notable breakdown between two historically intertwined tech power centers.

Musk's lawyers may have made a strategic trial error during Jared Birchall's testimony by opening a line of questioning about the acquisition bid, inadvertently giving OpenAI's counsel a legal entry point to admit new discovery evidence into the record.

Musk testified in a combative, visibly irritated style during Sam Altman's cross-examination, bristling at yes/no framing; Isaac flagged this as a potential liability with a jury that may weigh demeanor heavily.

Judge Rogers ruled that AI existential-risk and doomerism arguments are a sideshow distraction and has restricted how far Musk's side can take that narrative, undermining a key pillar of his framing.

The AI industry is beginning to internally acknowledge a serious public perception problem, and the high-profile OpenAI–Musk courtroom fight is compounding that damage at the worst possible time for the sector.

Non-sequestered jurors who can trade prediction markets tied to the trial outcome face a novel financial incentive for collusion, and Isaac argues court systems are not yet equipped to detect or prevent this.

The jury verdict in the OpenAI vs. Musk civil trial is advisory only; Judge Rogers retains full authority to disregard it and is solely responsible for setting any damages award.

Musk's team, with Ari Emanuel and others, submitted a bid to acquire OpenAI's for-profit arm; Musk's side claimed this was to establish a valuation baseline for the nonprofit, but OpenAI framed it as a legal harassment and disruption tactic.

OpenAI's legal team argued Musk's lawsuit is primarily competitive self-interest — that he filed because xAI is falling behind OpenAI, not out of genuine concern for the nonprofit mission.

Musk himself acknowledged in court that distilling from rivals' AI outputs is an 'open secret' across the industry, though Isaac notes this admission doesn't neutralize the legal exposure and may itself constitute fraud.

OpenAI's lead counsel argued Musk violated OpenAI's terms of service by using its model outputs to train xAI's competing products, framing him as a hypocrite who publicly warned about AI danger while secretly exploiting it for commercial gain.

CoreWeave dropped more than 4% on WSJ reporting about OpenAI's finances before rebounding 8.4%, signaling that AI infrastructure stocks are on a very short leash tied to hyperscaler capex confidence.

Meta has no AI diffusion problem: any model improvement to ad targeting can be AB-tested and deployed instantly across all apps, bypassing the enterprise change-management cycle that slows AI ROI everywhere else.

Meta's Family of Apps generated $30.8 billion in operating income last quarter, making the $6 billion Reality Labs loss financially immaterial and well worth maintaining as an option on the next computing platform.

Meta reached 3.58 billion DAUs and grew revenue 24% to nearly $60 billion in the quarter, with average ad prices rising 6%, demonstrating AI is already visibly monetizing in its core business.

AWS revenue consensus sits around $36.7 billion in the mid-20s% growth range, but the market is specifically hoping for a 30%+ print at some point in 2025 to validate the capex thesis.

Amazon's ads business grew 23% to $21.3 billion in the quarter, putting it on pace for nearly $100 billion annually — a quietly massive cash engine funding AI capex.