OpenAI’s GPT-4.5 just passed the toughest Turing Test ever—and didn’t just “pass.” It felt more human than the humans themselves.
In a blind test, judges picked GPT-4.5 as the human 73% of the time
Meta’s LLaMa-3.1 hit 56% — impressive
GPT-4o flopped with just 21%, near ELIZA
The riskiest way to use AI?
Turning it into a mirror for your own assumptions.
The most strategic?
Letting it question what you think is solid.
Next time it nods too quickly — pause and ask:
"What would the counterargument be?"
In 5 years, saying “I don’t use AI” will sound like
“I don’t use the Internet” in 2003,
or “I don’t use computers” in 1991.
The people experimenting now will have a massive head start over those still debating its limitations.
Stop debating. Start learning.
But most people aren’t ready to step into that space. Most systems are built for speed. But the next generation of leaders will be chosen by what they slow down for. This isn’t about the next tool.
It’s about what happens when machines do the work — and humans are left to lead.
AI is the new Wild West.
Everyone is rushing in. Tools evolve faster than teams can adopt them. The frameworks are unfinished, the consequences — largely unknown.
And yet, the race is on.
No one quite knows what we’re building — only that they must be first.(continue in comments)
And now that AI is generating dashboards and automating analysis — freeing people up for “meaningful conversations” — many leaders are discovering they have no idea what to say.
Suddenly, communication is not a soft skill. It’s the only skill that can’t be outsourced.
Every AI conversation is a mirror.
The quality of your outputs reveals the quality of your thinking in ways that are impossible to hide from yourself.
If the results feel shallow, the thinking was shallow.
When you get disappointing results, don’t blame the AI.
Fix your inputs.
Now who’s that?
LOL — a great damn manager.
And who’s a great manager?
The one who actually got the product built as close to the intended target as possible.
"Vibe coding" isn’t just blind building with AI prompts.
It’s breaking a complex mess into tiny, doable chunks
Explaining each part to ChatGPT like you’re onboarding and intern.
Debugging with the messy load of stuffs.
Then somehow stitching it all into a working product.
Great managers don’t just spot triggers.
They notice what lights people up.
A sudden spark in tone.
A topic they can’t stop refining.
A challenge they lean into.
When you pick up on these cues and build around them, you don’t just manage performance — you activate potential.
AI isn’t replacing decision-making.
It’s exposing where decisions were never structured to begin with. When you add AI, it needs inputs & criteria.
If those are missing, it runs on guesses.
People think they’re losing trust in AI.
But they’re losing trust in the system behind it.
@ToviGrossman I’m currently researching how AI intersects with emotional intelligence, judgment, and cognitive framing under uncertainty— especially in high-stakes decision environments. This kind of initiative feels deeply aligned.
@ToviGrossman It’s rare to see something that blends academic depth with venture-scale ambition so clearly. I’ll be following closely — and truly hoping this marks a shift in what’s possible here.
@ToviGrossman The brain drain in Canada’s tech and research ecosystem has been real, and painful to watch. Too often, some of the most thoughtful builders and researchers leave not because they want to — but because they can’t find the infrastructure here to match their ambition.
@camrrowe@yacineMTB@thescribblr Sounds great — please do share more!
I moved to Toronto during COVID, so I’m still catching up on what’s happening locally.
I’ve been deep in AI over the past few years (this is an older profile I’ve just revived — updates coming soon).