Groq

Uh, but if there's any sign of a slowdown, then, uh, uh, I think their pricing assumptions are gonna be a big struggle to, to maintain.

Well, and, and this is, this is a question I wanted to get your read on because w- with the whole, uh, Neo Cloud route that, that we've been seeing play out last, uh, 10 days specifically, you know, the, uh, one of the big ex- existential questions is, well, when the hyperscalers finish their own, uh, construction efforts on their own facilities, will this bridge capacity, w- you know, will it remain interesting to, um, the, the consumers of, of this compute? And, uh, SpaceX, SpaceXSI, whatever the company is called now, is in an interesting, uh, position because they have Cursor and they have Groq bot.

And the idea here is that we will be able to use some 50% of our capacity, was the number that he, that he threw out there last night, for our own, um, applications and stuff like that.

You know, this is a question that I, I don't think anyone has the answer to.

In, in, in, in that same regard, um, there's a post from Potato on Twitter.

I don't know, do you see, uh, she did a video and a post about shipping 2,500 PRs last month for Groq.

And, uh, she talked a lot about trust and automations being a part of it.

Like, how do you trust the output-

They give us business description, you know, email of, of POC, um, ICP-- rigorous ICP data, uh, target account lists, account-based selling lists from their CRM, and we, we match, um, we match, uh, enriched connections of connectors with, you know, ICP, uh, of business and target account list of business, and we show those matches in our platform and make it easy for, for connectors to facilitate double opt-in introductions that way.

Um, uh, we als- you can also use our product, uh, via our CLI in, in Groq, in, um, in, in Claude and, and I think OpenAI, uh, GPT.

Uh, I, I actually use it in Manus.

Um-

Uh-huh, yep, for sure.

That's probably outside of the interview playbooks and the, um, well, outside of the sales training and the certifications, I think the hottest debated topic at the moment in the Double Triple community is, do we go Claude, ChatGPT, Groq? Copilot's not even mentioned.

I-

It's just-

Pretty much any agent, to be fair, that can use a skill can work with this.

So Claude Code, Codex, Hermes Agent, OpenClaw, Cursor, Gemini, OpenCode, Kimi, Groq, and more.

So that's what it is.

Now the bigger question, do the numbers hold back? Because a claim like half the code is easy to make and hard to prove.

But anyway, so that's where I became fascinated for what, what this expert systems or this specialized systems can do for us.

So I always kept that in mind, and then when good old ChatGPT and Claude and, and Groq came into play, I really jumped into it several years ago because I think that they can do so much, so much for us.

And, and that's, that's, that's why I kept, um, uh, always, always trying to do something a-around AI.

Okay.

It's how those things become part of what IBM does, which is sort of the adopt and extend approach.

And I think one of the things we haven't identified yet is a moat to what? Like NVIDIA has a moat absolutely when it comes to training, but they bought Groq for a reason.

There are holes in its overall offense on owning enterprise inference, and what we haven't established is what enterprise inference is going to be shaped like.

The H100, which, uh, I will hold this to Jensen's feet as long as I can, he said last year at GTC that once, uh, Blackwell is up to speed, you won't be able to give away H100s.