Kevin Immanuel Gubbi(@kevingubbi):After a lot of discussions at SemiCon/tech week, it is evident that there's a long long road ahead to bring 'Superintelligence to everyone everywhere'. Chips, models, energy, physical AI, edge ai!? To anyone who's building in these areas and beyond... There's never been a better time to build... Keep going folks! (Goiing back to some key insights into how Agentic workloads use hardware I posted a while back.)

2026.10.15 00:15

After a lot of discussions at SemiCon/tech week, it is evident that there's a long long road ahead to bring 'Superintelligence to everyone everywhere'. Chips, models, energy, physical AI, edge ai!? To anyone who's building in these areas and beyond... There's never been a better time to build... Keep going folks! (Goiing back to some key insights into how Agentic workloads use hardware I posted a while back.)

Kevin Immanuel Gubbi@kevingubbi

2026.10.02 05:57

If AI agents spends more time waiting than generating answers, what should we actually be accelerating? And what does that mean for the next generation of AI chips? A look at where the time, memory and money go when agents get to work.