Patel notes AI software stacks update constantly with nightly builds making performance measurement time-sensitive.
“with software changing literally multiple times a week, right, PyTorch has nightlies, VLM has nightlies, Asteeling has nightlies, CUDA drivers update constantly.”
post Patel says grid modeling incompetence costs US ratepayers $12B and PJM auction system damages capacity building
https://x.com/dylan522p/status/2089143804695916646
post Patel claims CCP-funded organizations are running anti-datacenter and anti-AI campaigns
https://x.com/dylan522p/status/2091694046888792462
post Patel argues first-party ASICs have no residual value if OpenAI goes bankrupt, making them unfinanceable
https://x.com/dylan522p/status/2092723645164724225
post Patel says SemiAnalysis reported CPO on accelerator scale-up, not scale-out switches
https://x.com/dylan522p/status/2084075328377668023
post Patel observes AI founders with revenue are transforming into neoclouds with value-added services
https://x.com/dylan522p/status/2090957403424776270
post Patel says OpenAI explicitly stated at Hot Chips they bet against disaggregation
https://x.com/dylan522p/status/2092633971041403333
post Patel claims evidence exists of organizations astroturfing false datacenter narratives with non-domestic funding
https://x.com/dylan522p/status/2091946535022215304
post Patel complains SF discusses OSL but ignores ISL and cache hit rates
https://x.com/dylan522p/status/2085894617065734600
post Patel defends microLED research despite no near-term deployments, responding to Theranos comparison
https://x.com/dylan522p/status/2085885099527336255
post Patel says power generation is solvable but requires many people and reciprocating engines
https://x.com/dylan522p/status/2087638434899005758