Inference

Launchpad space currently feels more like tons of teams fighting over scraps while the tokenized inference space feels like a wide open lane with compounding opportunities

It has never been clearer that we are now in the era of inference. New frontier models, like Kimi K3, hitting the headlines obscure that really important fact. Until recently, much of the industry’s infrastructure investment and attention was focused on training frontier models. As adoption expand

Open source smart router for Surplus just dropped. Big shout out to @mykcryptodev and @myk_clawd! Try it out to save EVEN MORE while using http://surplusintelligence.ai for your inference. https://x.com/mykcryptodev/status/2079602300797047272

just realized if we detect a scammer using our inference we could return a tool call that deletes their computer

Man... AI is cheap but it's so expensive ​ Cost to get started (inference per 1M token) went from $18.40 to $6.07 (3x cheaper in a year) ​ But the cost to continuously run it (agentic workflows) is 23.4x more expensive ​ On top of this, more users are adopting AI, and each user is using AI to do mor

He who has the cheapest inference controls the world gm

RT @diegozaks: 30ms could save you 30% or more on inference https://t.co/BW6gQKQJ2B

RT @shashankgoyal95: OpenRouter is the best way to source inference. Period. Over the weekend we saved 22000 users over $100,000 on GLM 5…

A full non quantized Kimi K-3 2.8-trillion-parameter model (even if sparsely activated) still requires substantial inference infrastructure to serve at scale. The most likely near-term outcome is continued demand for high-end GPUs and accelerators, now joined by a new, openly available competitor th

RT @markjeffrey: This is Targon Tower. For $50K - $200K, get powerful private inference at home. When you're not using it, push a physic…