Your AI Assistant May Already Be a Better Work Spouse Than Your Human Colleagues

A lot of digital ink has been spilled recently about “Humanity’s Last Exam”, that somewhat dramatically-named test from the Center for AI Safety designed to probe the outer limits of artificial intelligence. Well, while the CAIS was busy collating closed-ended PhD-level questions that realistically only 0.000001% of humanity can answer, they missed something rather obvious: AI has already aced the most basic test of all – being a halfway decent coworker. And let's be honest, the bar for that was so low in most workplaces that it was practically a tripping hazard.

The evidence is becoming harder to ignore (much like Rajeesh’s unmuted heavy breathing and toilet breaks during Zoom meetings). While we humans were busy mastering the ability to do our shopping on Lazada whilst simultaneously maintaining strategic eye contact with the virtual meeting gallery, AI assistants swooped in and demonstrated what actual workplace competence looks like. No passive-aggressive emails, no performative pretense of having read the documentation, and miraculously, no strongly-worded suggestions about office air-con temperature settings.?

When you ask for that document from last quarter:?

• Human colleague: "Oh yeah, I think I remember that... let me check my emails... have you tried searching the shared drive? Maybe ask Kai Ling?"?
• AI colleague: "Here's the exact document, along with three related documents you aren’t aware of but are essential for background, and a summary of the conversations we’ve had about it since inception."

On meeting follow-ups:?

• Human colleague: sends the third "just circling back/gentle reminder" email, each one radiating increasing levels of passive-aggressive energy?
• AI colleague: "Here's an organized summary of action items pulled from our Slack, Teams and Zoom calls, synchronized with context, deadlines, and zero emotional manipulation."

On project feedback:?

• Human colleague: sits on your submission for three days, and finally replies with "looks good!"?
• AI colleague: "I've reviewed all 49 pages, cross-referenced against the client’s brand style guide, provided constructive feedback according to HR’s TGROW coaching framework, and identified three potential areas for improvement"

Is it slightly terrifying that our most reliable teammate doesn't even have a LinkedIn profile? Perhaps. But in an era where "per my last email" is considered a justifiable declaration of war, having a coworker that actually reads entire emails and documents before responding thoughtfully feels revolutionary.

We spent the last couple of years worried about AI replacing our jobs, when instead it's just making us humans look emotionally high-maintenance by comparison. It's setting workplace standards that humans can't match unless we develop photographic memory and the continuously-improving ability to make powerful inferences in seconds (my wild guess is that we won’t).

If you find yourself preferring your AI assistant's company to your human colleagues', know that you're not alone. You're just early to the realization that the future of work might be less about human connection and more about finding the perfect API integration for your ‘work spouse’ needs.

#AIatWork #ChatGPTo3 #DeepSeekR1 #QWen2.5Max #WorkSpouse