Prashanth (Manohar) Velidandi

AI Infrastructure Researcher | Co-founder at @InferXai , a Multi-Tenant Serverless Platform for Scalable Inference

[Prashanth (Manohar) Velidandi @PMV_InferX]

We just released Skill Function.

A new primitive in AI.

Not a prompt. Not a file. A callable AI endpoint that runs in the cloud.

Here’s how it works 🧵

[Prashanth (Manohar) Velidandi @PMV_InferX]

@AlexFinn Ya. This is getting some great reviews from our users too. Many saying better than Qwen 35b. Try it out : https://t.co/duuL04FKLI

[Prashanth (Manohar) Velidandi @PMV_InferX]

@ornith_ Surprisingly this is getting great reviews from our users. Some say , better than Qwen 35b level. try it out : https://t.co/duuL04FKLI

[InferX @InferXai]

Ornith-1.0-35B is now available on InferX.

If you’re looking to test a new 35B model with your own dedicated endpoint, you can spin it up now on InferX.

Get your endpoint

https://t.co/tZGoqNehm9

[Prashanth (Manohar) Velidandi @PMV_InferX]

@InferXai Try it out at https://t.co/duuL04FKLI

[Prashanth (Manohar) Velidandi @PMV_InferX]

@ornith_ Try it out at https://t.co/duuL04FKLI. It’s available now.

[Aravind Srinivas @AravSrinivas]

this is going to be the norm

[Prashanth (Manohar) Velidandi @PMV_InferX]

i think this is directionally right. interesting question is whether one giant ai should absorb every workflow, or whether expertise should be distributed into specialized capabilities. I think the latter scales better. let the primary agent orchestrate, and let specialists own their own context, reasoning, and lifecycle.

[Prashanth (Manohar) Velidandi @PMV_InferX]

@ivanburazin ‘When agents actually mature” - when/how is that?

[Prashanth (Manohar) Velidandi @PMV_InferX]

@abukinansafadi3 @jayair If they were the only two companies producing models, yes you’re very right.

[Prashanth (Manohar) Velidandi @PMV_InferX]

Indeed. i think the next step is making that tacit knowledge callable.

instead of every agent relearning or reloading the same domain expertise, workflows, and reasoning patterns, organizations will package them into specialized capabilities that can be reused across agents and models. that’s where i think the biggest leverage will come from.

[Prashanth (Manohar) Velidandi @PMV_InferX]

@ashugarg We’ve been saying this @inferx for a while. Ppl laughed when I said Anthropic doesn’t have a Moat. Here is why:

[Prashanth (Manohar) Velidandi @PMV_InferX]

On a recent episode of the @theallinpod , Bill Gurley @bgurley laid out what he called the "Dr. Frankenstein theory" regarding the frontier AI labs. After deep-diving into @AnthropicAI ’s public writings, including @DarioAmodei’s Machines of Loving Grace, Gurley pointed out an escalating, slightly eerie trend: these teams don't believe they are writing software anymore. They believe they are midwifing a centralized, superior digital deity.

At @InferXai , we completely agree with Gurley's skepticism. The industry has bought into a massive, highly synchronized delusion.

The idea that intelligence naturally scales into a single, omniscient, monolithic "God Model" defies everything we know about how intelligence actually works. Centering the future of technology around a single corporate oracle is an immense structural hazard, no matter how many 80-page "constitutions" or safety guardrails are layered on top of it.

Humanity didn’t build the modern world because one solitary super-genius emerged and the rest of the species blindly followed them. We built a civilization based on compounded, specialized, and distributed intelligence. True capability scales through high participation, friction, redundancy, and a collective network of diverse actors working in parallel.

AI will scale exactly the same way. The future is not a trillion-dollar model sitting on a mountain. it is a highly coordinated, deeply integrated collective.
And here is our core conviction: The future of intelligence is not model-centric.

A model is just a runtime. The real breakthrough isn't the single asset; it’s the orchestration fabric that binds a distributed civilization of compute together. What the industry will ultimately call this interconnected web remains to be decided.

Internally at InferX, we already know the answer. I’ll talk more about it soon.

[Prashanth (Manohar) Velidandi @PMV_InferX]

absolutely. the long-term risk is concentrating all of a company’s knowledge, memory, and workflows into a single agent owned by a single vendor. i think the future is more modular. agents should be replaceable. models should be replaceable. expertise should be portable.

that’s part of why i’m interested in @SkillFunction instead of one giant agent that knows everything, specialized skills own the expertise and can run on different models and different runtimes. the agent becomes a thin coordinator rather than the place where all organizational knowledge lives. https://t.co/l0RW3KZe7f

[Prashanth (Manohar) Velidandi @PMV_InferX]

@karpathy @SkillFunction can do exactly this with your own intelligence docs , and with much smaller models. You don’t need Opus-level compute as long as the skill has the right context and the right model bound to it. https://t.co/4HucyJjwJQ

[Prashanth (Manohar) Velidandi @PMV_InferX]

@AnthropicAI just launched @Claude Tag. Agents in Slack with persistent memory and real tasks.

This is exactly the direction we’ve been building toward.

Callable expertise. Persistent knowledge. Serverless inference.

The market is catching up.

Can’t be more exciting to be building the future of AI right now.

More here👉🏼: https://t.co/q3msXH0Esj

[Claude @claudeai]

Introducing Claude Tag, a new way for teams to work with Claude.

In Slack, Claude joins as a team member with access to the channels and tools you choose. Tag Claude in and delegate tasks to it while you focus on other work.

[Skill Function @SkillFunction]

Skills are coming to every platform. Microsoft in Excel. Google in Gemini Enterprise. But they’re all closed, locked to one vendor, one model, one ecosystem. Skill Function is the open infrastructure layer - callable from any agent, any model, any platform.

Own your intelligence.

Try it for free: https://t.co/OhGiFipUcL

[Prashanth (Manohar) Velidandi @PMV_InferX]

@satyanadella Looks like Skills are becoming the new apps. Microsoft in Excel, Google in Gemini Enterprise, but all closed to their own ecosystems. @SkillFunction is the open infrastructure layer .callable from any agent, any model, any platform. Try it for free: https://t.co/b6nNNeX6Gm

[Prashanth (Manohar) Velidandi @PMV_InferX]

@pbakaus Would love to make this native to @SkillFunction

[Prashanth (Manohar) Velidandi @PMV_InferX]

@sarahwooders MCP and skills solve different things. MCP is how agents discover and call tools. Skills are what those tools actually do. They work together . @SkillFunction uses MCP for discovery but the execution happens in isolated cloud contexts. You need both.

[The Information @theinformation]

Some Anthropic and OpenAI customers are cutting AI costs by 90% or more as cheaper models prove good enough for more tasks.

Full story: https://t.co/WTtFt9CP2N