Working with AI/ML since 2015. Enterprise AI implementation & SME advisory. Writing for people building AI systems and leaders deciding where to invest.
QLD, Australia & Asia
Joined October 2013
- Tweets85.4K
- Following350
- Followers51.7K
- Likes27.1K
Pinned Tweet
An AI demo proves a capability. Production tests the economics, reliability and organisation behind it.
I write about those decisions: agents, engineering leadership and enterprise AI.
Building systems or leading adoption? Follow along and bring your experience.
2
2
1,145
One of the best things happening right now is that people who rose through the ranks through corporate politics are now losing ground, because the new AI-driven environment exposes those who don't deliver.
They're living on borrowed time. I know, because that's the environment I operate in.
I am very happy with the direction we are moving in.
3
682
With the current tooling and workflow, if you see more than 4-5 people in a meeting, they're most likely doing it wrong.
The volume and sophistication of cyberattacks that organisations are experiencing today are reaching extreme levels.
And things will only get worse if cybersecurity teams continue approaching these threats with conventional methods.
The best defence is increasingly a good offence.
Organisations need to move beyond reactive security and start deploying autonomous AI agents that proactively hunt threats, identify vulnerabilities, and counter attacks at machine speed.
The most likely superintelligence takeover scenario is one in which we never recognise that a takeover has occurred and simply continue living our lives as if nothing has changed.
Technological capability is not the same as commercial value.
The real money is in building agentic frameworks that can save enterprises tens of millions of dollars by fundamentally changing how they operate. Technology is just a part of that challenge.
Just because a new model allows you to do something you couldn’t do yesterday doesn’t mean you’ve created business value.
The value comes from understanding your business context and knowing how to apply that new capability to solve a real problem.
This is amazing.
In one of my favorite experiments, I told Fable 5.2 to build an endless, hands-off journey through 3D worlds linked by portals.
25,000 agents created over 450 worlds inspired by books, art, places, all in code.
Play: portals-endless-worlds.netli…
Github: github.com/emollick/ai-porta…
1,364
Interesting insider view of Revolut’s pivot to free-weight models:
They’ve cut token costs by 90%, developers have much more freedom to use tokens, and Claude/GPT are now the exception rather than the default.
What makes this particularly interesting is that Revolut is a bank operating in Europe, so this is happening in a highly regulated industry within a highly regulated region.
But the biggest takeaway here is: don’t scale first, tune first. Apparently, 80% of the gains came from hyperparameter tuning, not adding more GPUs.
1
3
1,404
Billion-dollar question: What product should we sell to people who will get tired of AI and yearn for the old days?
5
8
2,653
I've seen CEOs fired for far less.
Breaking news from @ianmking: Nvidia CEO Jensen Huang admits to extramarital affair and the company board is now investigating him being targeted by a Chinese espionage plot. He remains CEO for now. bloomberg.com/news/articles/…
7
1,717
I have been working with AI systems since 2015. My first AI-related article was published in 2017. These were before transformers.
I ran AI and data transformation programs worth more than $ 100 million.
Today, a 30-year-old salesperson who became an AI consultant last year told me I should read a bit more about AI.
Nature is full of miracles.