Doug O'Laughlin

Jon Y × Doug O'Laughlin × Jordan Nanos

Google’s talent drain, including Jeff Dean, John Jumper, Noam Shazeer, and David Silver, raises questions about whether system-level judgment can be replaced by more compute.The risk is execution, not earnings: Google may remain highly profitable and strong in TPUs while quietly losing frontier-model leadership, as agentic coding accelerates demand for bespoke software and Terafab faces a heroic physical ramp.

Jordan Nanos × Doug O'Laughlin

Kimi K3 is now a clear top-three model by benchmark composites, while Dylan ranks it second for practical use as Opus access remains frustrating and restricted.Its 2.8-trillion-parameter scale requires B300, GB300, or MI355X-class hardware, but $3/$15 per million input/output tokens suggests attractive economics; adoption, Western sovereign demand, and routing layers remain the key commercial variables.

Jordan Nanos × Dylan Patel × Doug O'Laughlin × Max Kan

GPT-5.5 brings OpenAI back into the frontier conversation after Anthropic surpassed it on a like-for-like revenue basis, but Claude 4.7’s quality advantage over 4.6 remains unproven despite a 6x fast-mode premium.Token costs are beginning to pressure heavy users as new tasks drive Jevons-style consumption, while China’s compute constraints widen the open-source gap and revive the CLI-versus-app battle over agent orchestration.

Jordan Nanos × Doug O'Laughlin × Myron Xie × Copper Wei (Wega) × Howie

Rubin’s adaptive compression engine makes sparse FP4 potentially usable, delivering up to 50 PF effective performance from 35 PF dense processing, while HBM4 targets 22 TB/s per chip and may bin suppliers by capability.Cableless compute trays and NVIDIA-controlled SoCAMM procurement address GB200’s yield pain and memory tightness, but 2.3 kW GPUs and a likely 1H27 mass deployment leave ramp hiccups and HBM pricing as risks.

swyx (Shawn) × Doug O'Laughlin

Claude Code with Opus 4.5 now behaves like a junior analyst, compressing PhD-scale projects into days while still requiring expert review of its frequent mistakes.Adoption near 4–5% of public GitHub and a memory shortage that could lift DRAM prices another 100% point to accelerating agent demand, constrained infrastructure, and exposed information-work roles.

Cam Quilici × Bryan Shan × Doug O'Laughlin × Jordan Nanos

InferenceX 2.0 shows GB200/GB300 delivering 20x DeepSeek-R1 throughput per GPU versus a fully tuned H100 at low interactivity, and 80–100x at 100 tokens/sec/user.NVLink’s 72-GPU domain, software optimization, and multi-token prediction drive the gap, while MI355’s roughly 25% TCO advantage over B200 highlights margin potential; composability, legacy fleets, and larger frontier models remain risks.

Jordan Nanos × Doug O'Laughlin

Claude Code now authors roughly 4.7%–4.8% of public GitHub commits, while Opus Fast and GPT-5.3-Codex-Spark test how much users will pay for faster inference above the $2,400/year Max tier.Memory has shifted from the worst downturn since 1996 to a multi-year shortage as HBM absorbs wafer capacity, CPUs tighten, and even new supply may not close the gap; storage demand hinges on logs, synthetic data, and generated video.