Rishi (@notnotrishi) 's Twitter Profile
Rishi

@notnotrishi

curr: ⚒️ with ai | past: early @beeper, dropout @mila_quebec, among other things

ID: 184912642

calendar_today30-08-2010 20:01:32

1,1K Tweet

217 Followers

50 Following

Rishi (@notnotrishi) 's Twitter Profile Photo

My AI usage pattern completely switched over the last one month… a lot more of GPT5-Codex (over Sonnet) for coding tasks …and Claude (over ChatGPT) for general tasks!

Rishi (@notnotrishi) 's Twitter Profile Photo

Hmm it ran when I was asleep but unfortunately stalled last min and doesn't have the last checkpoints Might tweak a few things and rent some GPUs to run it over the weekend again, but the best part is I could train a simple example on my Mac before I decide to rent GPUs

Hmm it ran when I was asleep but unfortunately stalled last min and doesn't have the last checkpoints

Might tweak a few things and rent some GPUs to run it over the weekend again, but the best part is I could train a simple example on my Mac before I decide to rent GPUs
Jon Saad-Falcon (@jonsaadfalcon) 's Twitter Profile Photo

Data centers dominate AI, but they're hitting physical limits. What if the future of AI isn't just bigger data centers, but local intelligence in our hands? The viability of local AI depends on intelligence efficiency. To measure this, we propose intelligence per watt (IPW):

Data centers dominate AI, but they're hitting physical limits. What if the future of AI isn't just bigger data centers, but local intelligence in our hands?

The viability of local AI depends on intelligence efficiency. To measure this, we propose intelligence per watt (IPW):
Rishi (@notnotrishi) 's Twitter Profile Photo

Brain dump on building AI products: 1/ It is easy to ‘vibe code’ an AI product today, but the harder part is making it reliable. Don't think much has changed -- the burden/effort has now shifted to R&D. One small wrong assumption can send you back to zero. This is not new --

Tomasz Tunguz (@ttunguz) 's Twitter Profile Photo

70% of production AI teams use open source models. 72.5% connect agents to databases, not chat interfaces. This is what 375 technical builders actually ship - & it looks nothing like Twitter AI. 70% of teams use open source models in some capacity. 48% describe their strategy as

70% of production AI teams use open source models. 72.5% connect agents to databases, not chat interfaces. This is what 375 technical builders actually ship - & it looks nothing like Twitter AI.

70% of teams use open source models in some capacity. 48% describe their strategy as