ceo @pipeshiftai | yc s24 | forbes 30u30
San Francisco, CA
Joined August 2026
- Tweets111
- Following206
- Followers343
- Likes130
flew to bangalore after 2+ years to meet some of our amazing customers and i immediately realized why founders leave.
- sat in an uber for 1.5hr to travel 8 miles
- touted to be a tech city, but internet connectivity is sh*t in office zones like indiranagar
- tried walking over to meetings to avoid traffic and found no walkable area, my sneakers are ruined
not able to do perform even half as efficiently as compared to when in sf.
it’s just not worth it. building something is already hard enough. building it here makes it 10x harder.
going oss is a no brainer.
fire anyone in your team who argues otherwise.
Own your intelligence.
Anyone that opposes this has something to sell you (perhaps your own intelligence back to you.)
Local models need to continue proliferating. The “absolute unit” massive models may be out of reach for now locally, so your best option is to, indeed, keep your options open. Own your harness, data, connections, runtime: your agents. Embrace open source.
bs bait. either he has shady providers or smokin something cheap.
you’d take those GPUs yourself before dunking on the market if this was legit.
we work with enough verified colos and providers who buy these GPUs on their balance sheet.
none of this is true.
and if you’re a provider selling compute at $4.6/hr w/o commits, DM me right now. we’re taking it all.
this is why open source will win.
closed model labs are hypocritical at their core.
Shame on @DarioAmodei / Anthropic. I never distilled Claude—rollouts came from a strong open-source model; Claude just helped submit jobs and run analysis. Still got tagged for "training a competitor," and now anything touching RFT, SFT, or training larger models gets me blocked, Opus 5.5 downgraded to Opus 5, or the session killed outright. You built your models on the world's data, then treat paying users doing research automation as the enemy. Pathetic.
i’ve seen priced rounds being done for 2cr in India.
the sheer lack of value for time, not just for the founder, but even themselves, is outrageous.
love watching people flip their opinions on compute demand off a few tweets from a guy who just wants neocloud stocks tanking to corner all compute.
same guy claims compute pays itself in 6mos.
funny how the loudest "stranded compute" voice is also building more of it than almost anyone. xai + spacex alone put him at ~10gw, and some reporting says it's closer to 20.
so no, he's not an impartial observer. every competitor that doesn't survive this cycle is one less in his way.
no doubt humans are 99% chimp dna.
the red graphs are the proof. all sheep.
Pretty much.
As of the past few weeks, debt markets have almost entirely shut off access to money for any company that doesn’t have substantial revenue independent of AI data centers.
SpaceX closed $40B of debt for GPUs, but we’re investment grade and have Starlink & launch revenue.
That means the 200+ neoclouds are going to see guppy-level death and consolidation. Some will make it through, but not many.
i wish more folks talked about this
waiting for closed model companies to cry for regulations.
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model.
🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal.
🔹 Full-project coding: games in Unreal, Unity & Godot.
🔹 Excels at parametric 3D modeling in Blender & CAD.
🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%.
🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks.
Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi.ai, Kimi Work, Kimi Code, and the Kimi API.
Open weights by Nov 4, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
forget 3mo gpu rentals.
buy a dogpatch apartment and offer it batch-to-batch to yc founders.
best rental yields globally.
