It's Time To Name And Shame - WAN Show June 2, 2023

Episode Summary

Executive Summary: A long, meandering WAN Show from Computex in Taiwan focused on how the tech industry is shifting away from consumer gaming hardware toward data center/AI, why Nvidia’s gaming side feels increasingly sidelined, and how reviewers and creators should maintain independent viewpoints. The episode also covered the Dolphin Steam launch delay, Reddit API pricing backlash, a Gigabyte firmware backdoor, Apple headset expectations, and Linus’s decision to call out a pool contractor and a Kickstarter-backed case company for unethical behavior.

Main Topics: Computex reveals a data-center-first industry shift (Priority: 5/5): The hosts argue that the most exciting innovation at Computex was in data center, AI, cooling, and mobile/accelerator tech rather than consumer gaming hardware. They note the near absence of major consumer desktop GPU/CPU news and question whether these developments will meaningfully trickle down to gamers. Nvidia’s business incentives and GeForce’s future (Priority: 5/5): A major segment explores Nvidia’s skyrocketing valuation, the 4060 backlash, and the idea that GeForce may be strategically constrained by Nvidia’s larger AI/data-center priorities. Linus even speculates that GeForce could be spun off to better serve gamers without shareholder pressure. Ethics, accountability, and naming/shaming companies (Priority: 5/5): Linus details why he publicly called out a pool contractor for repeated delays and why he refused to cover a case maker/Kickstarter-style product that he felt had effectively bait-and-switched backers. The discussion centers on integrity, refunds, and consumer trust. Platform control and creator independence (Priority: 4/5): The show discusses how reviewers should not all converge on the same opinion, why different outlets should use different approaches, and how brands try to manipulate creators through access, sponsorship, or pressure. This includes a clarification of Linus’s conflict with Nvidia and the importance of critical thinking. Security and software ecosystem risks (Priority: 4/5): The hosts cover the Gigabyte motherboard firmware issue, where a firmware updater could drop and run an executable from disk, and the Dolphin Steam postponement after a Nintendo DMCA threat. Both stories illustrate how platform control, firmware complexity, and legal pressure can affect users. AI, generative content, and gaming (Priority: 4/5): The episode debates whether AI will enrich games or simply be slapped into products as a marketing gimmick. The consensus is that AI will be used broadly, but many early implementations will be clumsy or cost-cutting rather than genuinely useful. LTX 2023 and community-building (Priority: 3/5): The show ends with a detailed promo for LTX 2023, including creator attendance, workshops, hands-on activities, and a push for volunteer signups and ticket sales.

Key Arguments: Consumer-facing gaming hardware is getting less attention because the most meaningful R&D and business incentives are now in AI, networking, and data-center accelerators. PCIe Gen 5 and faster SSDs matter far more for simplifying board design and enterprise efficiency than for gaming performance, so consumer adoption may remain limited. Nvidia’s GeForce division may be structurally disadvantaged because gaming margins look weak compared with AI/data-center products, even though the GeForce team still appears genuinely passionate about gaming. If hardware makers don’t need to compete, they tend to stagnate or raise prices; competition among Nvidia, AMD, and Intel is still necessary for consumers. Reviewers should be free to disagree; forcing all outlets to repeat the same take creates dangerous groupthink and reduces useful consumer context. Brand relationships are routinely shaped by access, incentives, and manipulation; creators must keep their independence and not confuse friendliness with trustworthiness. Companies that take consumer money and fail to deliver should provide refunds; offering only a credit toward a more expensive product is not sufficient ethical remediation. AI in games is likely to begin as overhyped, sloppy, or management-driven cost cutting, then later settle into more subtle and useful applications. Dolphin’s Steam launch delay shows how quickly platform holders and copyright owners can shut down even legitimate emulator distribution when proprietary keys and DMCA concerns are involved. Reddit’s API pricing appears intended less to monetize fairly and more to crush third-party apps and limit competing access paths, especially as AI scraping becomes more valuable.

Data Points: Nvidia valuation: briefly reached $1 trillion - After strong AI demand and alongside controversy around the RTX 4060 launch RTX 4060 reaction: stock jumped about 25% the following day - Referenced as a contrast between gamer backlash and investor enthusiasm Gigabyte vulnerable models: at least 271 motherboard models - Firmware backdoor issue affecting recent Z790 and X670 boards as well Pool project timeline: 2 years - Linus says the pool/landscaping project has been stalled for two years Pool project remaining work: a couple weeks - He says the job was down to roughly two weeks of work for months Computex focus: almost no consumer desktop news - Observation that Intel, AMD, and Nvidia showed little for consumer desktop Reddit API pricing: $12,000 per 50 million requests - Apollo developer’s estimate of Reddit’s proposed pricing Apollo request volume: 7–10 billion requests/month - Used to estimate monthly cost under Reddit’s pricing Apollo estimated monthly cost: about $1.7 million/month - Based on Reddit’s API pricing and Apollo usage Apollo per-user requests: 344 requests/day - Average usage cited for Apollo users Apollo user cost: about $2.50/month - Per-user cost implied by Reddit’s pricing Imgur comparison: $166 for same number of requests - Used as a benchmark showing Reddit’s price is far higher Twitter API comparison: 10,000 requests for $100 - Mentioned as another example of expensive API access Quest 3 price: $499 - Meta Quest 3 announced for fall release Quest 2 price drop: $299 - 128GB Quest 2 starting June 4 LTX 2023 ticket sales: over 3,500 tickets - Sales count mentioned during the promo segment Microblade server density: 8 full computers in a 3U rack - Supermicro Ryzen-based microblade server example Nvidia network card compute cores: 16 core CPU on the card - Used rhetorically to show how much compute is being embedded in infrastructure products

Pivotal Quotes: "I think they should spin off GeForce." — Linus: Argument that Nvidia’s gaming division is being structurally hampered by the company’s AI/data-center incentives "If you take someone’s money and don’t ship them a product then you should give them their money back." — Linus: Discussion of the passive case/Kickstarter controversy and why a voucher was not enough "We don’t all have to say the same thing." — Linus: Defense of reviewer independence and criticism of audience pressure toward groupthink

Implications: Listeners are encouraged to be more skeptical of brands, platform holders, and consensus reviews. The industry appears to be drifting toward AI/data-center priorities, so gamers may need to temper expectations for trickle-down innovation.