CTO Turned Advisor | Helping Vendors Resonate and IT Leaders Execute. Engage with my virtual twin https://nitter.cf/t.co/7fh1X8hbEJ. Independent Advisor.
Chicago, IL
Joined June 2009
- Tweets86.3K
- Following505
- Followers27.2K
- Likes21.8K
I'm at capacity and feel like I shouldn't take on new clients for a couple of months. However, that's exactly the point: you sell more. Because of the way execution works, it's a queue. And if you wait until the queue is empty to sell, you leave capacity on the floor.
The Illinois Department of Revenue and IRS are incredibly helpful when you are trying to figure out a way to send them money.
And I’m not even being sarcastic. The customer support is pretty good.
Why is this the part that works?
This goes for stress testing any idea that’s not in the training data of any LLM.
If you give it a pattern it hasn’t seen, it struggles to add value.
The same when you give it an anti-pattern as a task. “Forget your training” isn’t a thing.
Anthropic vs. Google. Or better: Haiku vs. Gemma 4- which one interprets "yes" better in a customer service call?
That may be the perfect example of how we ask the wrong questions in agentic AI.
ctoadvisor.substack.com/p/wh…
What's the NeoCloud risk for CTO's? In the latest @thectoadvisor, @calvinhp and I chat about the impact of the uncertainty in the NeoCloud market on enterprise AI plans.
I don't think @GoogleCloud realized that Agent Compute Engine may be more of an economic argument to migrate from @VMware to GCP than it is an Agentic Compute platform.
ctoadvisor.substack.com/p/go…
Today is the last day to return my M5 Ultra Studio with 256GB. I wanted to hold out for the 512GB system and I got my answer. My sales rep emailed me. And I have to say, "OUCH!!!"
I'm not willing to pay this much for 256GB more RAM. It's a $15,500 option. That's Crazy!
For my lab, I could double the number of Studios and RAM for the same cost by simply dropping the 2nd machine to a 1TB drive. Double the power usage as well.
I'd double my aggregate throughput. This is where the Mac falls short compared to NVIDIA GPUs.
Knerd.AI has been my side project to discover the moat. How do I take all of my AI research and turn it into a company that solves the core challenges of realizing the advantages of AI?
Software is an enabler but no longer a moat.
Everyone can do software. Or at least think they can.
Today the real moat is boots on the ground. But that doesn’t scale. At least not in a way investors want to see.
What a time to be an entrepreneur.
I have a pretty good relationship with @GoogleCloud. The closest I've gotten to a TPU was the LEGO kit from Cloud Next.
Yet Oracle Cloud gets TPUs?!?
@rseroter and @ballen_clt I thought we were friends!!!
🤣
