Did you know that we have LinkedIn, Instagram and X accounts
that you can follow?

Hi {{first_name|everyone}},

IBM’s CEO says the smartest AI model isn’t always worth paying 10x more for, and we’re looking at why more businesses are starting to agree. In the Buzz, someone used Claude Code to dig through NASA data and may have found a planet nobody knew existed, and there’s a look at why keeping track of multiple AI agents is still such a headache. Plus there are some new tools to try and a 3D-printed design project that’s fun to look at, so let’s dig in!

🤿 DEEP DIVE

IBM's CEO has a pretty simple argument about AI costs.

Apparently, having the smartest AI model isn't necessarily worth paying 10x more for it.

That's basically the argument IBM CEO Arvind Krishna made during the company's Q3 2026 earnings call. And it's a pretty interesting way to look at where enterprise AI spending is heading.

Krishna says businesses are becoming much more conscious of what they're actually getting for their AI budgets. Especially when those budgets start getting out of control without much return to show for it.

And the numbers he used to explain this are worth looking at.

Back in 2024, Krishna said there was a pretty big performance difference between the best frontier AI models and cheaper alternatives.

On a broad benchmark, the top model might score 70 while an alternative scored just 20 or 30. That's a pretty hard compromise to justify, even if the cheaper model costs less.

But by late 2025, that gap had narrowed. Instead of 70 versus 20, it was more like 70 versus 50.

And Krishna believes the models are moving toward a point where the best might score 100 while an alternative scores 90.

Which brings us to his actual point.

If a model can do 90% of what the best one does, but costs just 10% as much, why would a business automatically choose the expensive one?

That's the calculation he's seeing more companies make.

There's another part to this that makes the argument more interesting.

Krishna says a typical large enterprise could run these open-weight models using just two or three racks of infrastructure, bringing token costs down to somewhere between 10% and 25% of what frontier models cost.

And because they're open-weight, companies can run them on their own hardware or private clouds.

That matters for businesses dealing with sensitive information.

Krishna pointed out that using public AI models can leave traces of company activity in logs, including what kinds of questions people ask and potentially elements of the data involved.

Running models privately gives businesses control over those logs and their data.

So it's not just about saving money. It's also about deciding where the AI runs and who controls the information around it.

Krishna isn't saying frontier models are suddenly unnecessary. He acknowledges that some tasks still benefit from the more expensive options.

But his argument is that most enterprise tasks don't necessarily need that extra performance.

And that's the part worth remembering. The question businesses are asking isn't just which AI model performs best anymore.

It's whether the difference in performance is actually worth the difference in price.

🤝 FROM OUR FRIENDS

A quick share we thought you might find useful

Your client doesn't care how their site was built. They care that it launches on time, runs on their own domain, and keeps working long after you hand it over.

With Lovable, you describe what your client needs and get a working version you can show them in the first meeting. Then you take it all the way to a live site on your client's domain, built to be found on Google. Lovable's project monitoring feature spots bugs and suggests fixes, so your name stays on work you're proud of.

Start today and take $5 off your first month.

Build for clients, $5 off

_{^{Discount of $5 applies to first month of any paid plan. You need to sign up via this webpage in order to claim the discount. The discount applies to new and returning users without an active paid plan. Terms apply.}}

🔥 WHAT’S THE BUZZ

What people are building, sharing, and talking about

Anyone managing several AI agents knows the frustration of wondering whether something is still running or waiting for attention. Better status reporting could make agent dashboards, notifications and automated workflows much easier to manage.

Pavel used Claude Code to search NASA data and may have found a planet nobody knew existed.

An interesting example of AI making specialized research more accessible. Someone with curiosity, publicly available data and the right tools can now investigate questions that previously required substantial programming expertise. The important lesson is how Rabtsevich used independent checks to challenge the AI's findings.

⚒ TOOL SNAPSHOTS

Futuristic tools within AI, no-code, and productivity

🤖 AI MADE THIS

Interesting and inspiring creations made with AI

LUCID - AI-generated forms you can 3D print and reshape

LUCID uses AI to help designers create strange 3D forms and then turn them into real, printed objects. The clever part is that the AI doesn't just hand over a finished shape, designers can see and change the logic producing it. The project is currently being shown at Dutch Design Week.

AI used:
The project confirms AI-assisted / agentic generative design workflows, but did not specify the underlying foundation model.

ℹ️ ABOUT US

The Intelligent Worker helps you to be more productive at work with AI, automation, no-code, and other technologies.

We like real, practical, and tangible use-cases and hate hand-wavy, theoretical, and abstract concepts that don’t drive real-world outcomes.

Our mission is to empower individuals, boost their productivity, and future-proof their careers.

We read all your comments - please provide your feedback!