AI did not kill Chess
#148 - Why the truth about AI lies between the doom and the hype
Hello friends,
I hope you had a great week! As we’re enjoying summer holidays on the beach in Italy, I stumbled across a tweet that made me (once again) reflect on the role of AI.
A lot of people keep talking about how AI is going to disrupt ABSOLUTELY EVERYTHING it touches, even more so when new models launch (like ChatGPT 5 this week).
The tweet was reacting to someone who claimed that, once the engines solved the Chess game, human grandmasters would contribute “nothing but noise.”
That tweet got me curious. So before digging deeper, I spent a few minutes looking for other examples where the “AI will kill us all” conclusion turns out to be premature. I wanted to find moments where working with AI can actually make things more engaging and fun. Since it’s summer, I’m keeping the tone light and lining up stories from across games, creativity, and work that illustrate how human plus AI often beats doom‑and‑gloom.
Games: Deep Blue to Move 78
When IBM’s Deep Blue edged out Garry Kasparov in 1997, analysts queued up to write chess’s obituary. Instead, the game went mainstream. Today Chess.com reports more than 140 million registered accounts, up from about 24 million a decade ago. The 2024 Candidates Tournament streamed to a peak audience above 2 million concurrent viewers, and world-title sponsors now pay seven-figure sums for logo space. Forbes notes that modern chess is “far more popular than ever before,” with new champions like Ding Liren proving the talent pipeline is global, not niche.
Chess has, litterally, never been more popular.
Engines didn’t kill the board—they became its relentless personal trainers. A kid in Lagos or Lima can boot up Stockfish on a phone and spar at a level that would have humbled Bobby Fischer. Average club ratings have crept upward; even openings once considered worthless get re-examined because the computer found gold two moves deeper.
Go followed the same arc, only faster. Benjamín Labatut’s The MANIAC closes on the 2016 AlphaGo–Lee Sedol match. Down three games, Sedol unleashed a shoulder-hit—Move 78—that the commentators labeled insane. AlphaGo’s win rate, displayed in cold percentages, plummeted in real time; the machine resigned. Sedol later said that single victory “was enough.” The Seoul crowd agreed, roaring as if humanity had stolen back a sliver of the divine.
What happened after the applause matters more. Researchers combing through 5.8 million professional moves discovered a measurable jump in Go “novelty” and decision quality post-AlphaGo. Top players began mixing AI-generated lines into their repertoires; mid-rank pros improved faster because cloud engines made elite analysis a commodity. The tactical summit belongs to silicon, but humans pushed the strategic frontier outward.
The pattern is hard to miss: first comes dread, then a machine victory, then a bigger, weirder, more crowded scene. Engines expose our blind spots; we respond by leveling up. If history keeps rhyming, the next breakthrough won’t end the game—it’ll hand us a sharper tool and a steeper learning curve.
Culture & Commerce: DJs, Coffee and Counter-Strike
Spotify can auto-mix four hours of EDM without breaking a sweat, yet clubs still shell out for human DJs. According to the latest U.S. Bureau of Labor Statistics release, radio announcers plus live-event DJs make up roughly 47,000 jobs, and the market still expects about 4,800 openings every year through 2033. That’s flat growth, not extinction. The reason is obvious to anyone who’s closed a dance floor: the crowd wants a story, not a shuffle. A good selector reads the room, nudges tempo, holds back a drop an extra bar to make strangers cheer in unison—skills today’s recommendation engines still fake poorly.
Home espresso machines tell a similar story. De’Longhi, Breville, and robot baristas can pull a clean shot at the push of a button, yet America’s specialty-coffee segment is projected to reach $81.8 billion by 2030, compounding at 9.5 percent a year. World Coffee Portal expects branded shop counts to pass 51,000 outlets by 2029. People aren’t paying €4 for 30 ml of liquid; they’re paying for the lo-fi playlist, the latte art, and the option to linger with a laptop while someone else handles the cleanup.
Gaming? Same pattern. Counter-Strike has battled aimbots and wall-hacks for two decades, but Global Offensive (and now Counter-Strike 2) still attracts around 11 million monthly players. In March 2023 the game set a new concurrent-player record of 1.8 million—impossible if cheating had truly ruined the fun. Valve’s anti-cheat arms race, league rulebooks, and stream overlays are the proof the community values fair competition enough to keep fighting for it.
Across dance floors, coffee counters, and virtual bombsites, technology keeps raising the execution ceiling while leaving the vibe—curation, hospitality, fair play—squarely in human hands. Customers notice, and they keep showing up with wallets open. The mechanic may be automated, but the meaning remains delightfully analog.
Sport & Spectacle: VAR and the Game We Still Watch
Few technologies have been more divisive in sport than VAR. The video assistant referee system promised to eliminate “clear and obvious” errors in football. And in a narrow, technical sense, it has. Offside lines are drawn to the millimetre, handballs slowed to 240 frames per second, goals overturned when the ball is shown to be a hair over the line.
But has VAR made football less watchable? Not really. Stadiums are still full, TV rights keep breaking records, and the World Cup final remains one of the most-watched events on earth. Fans still argue about borderline calls—only now the freeze-frame is part of the ritual. Players adjust, knowing that a tug in the box might be spotted from a dozen angles. Managers adjust, crafting set-pieces that survive forensic replay.
Like chess engines or coffee-making robots, VAR changes the mechanics but leaves the meaning intact. The thrill still comes from a ball curling into the top corner, a last-ditch tackle, or the noise of 60,000 people erupting in unison. Technology can tidy up the rulebook; it can’t script the drama.
Work & Productivity: Augment, Don’t Automate
The dread that AI will erase whole career ladders echoes every few months, but the numbers point in a subtler direction. A Washington Post analysis mapped one million Anthropic chat sessions onto 700 occupations and found that by the end of 2024, AI was touching about 25 percent of the average job’s daily tasks—yet almost always on the augment side. For educators and librarians, roughly 40 percent of duties can be enhanced by AI, while only 23 percent of tasks in computer-and-math roles are ripe for outright automation. The pattern is less “robots take your desk” and more “robots take your busywork.”
That lines up with Garry Kasparov’s hindsight on chess software: “Humans who embraced machines replaced humans who didn’t.” The winners aren’t the fastest typists or the deepest subject-matter experts; they’re the quickest adapters—people who offload the rote parts to an LLM and then focus on the judgment calls. Nvidia’s Jensen Huang frames prompting as a high-level cognitive skill: ask sharper questions, get sharper leverage.
Magnus & ChatGPT
One final anecdote that brings this full circle: on July 10, 2025, Magnus Carlsen – arguably the best human chess player in history – posted a screenshot of a game he’d played against ChatGPT while traveling. The result? Carlsen forced ChatGPT to resign in 53 moves, without losing a single piece. ChatGPT conceded, calling the play “methodical, clean, and sharp,” even though it had misread Carlsen’s skill (estimating him as a 1800–2000 FIDE player, when his actual rating is 2839) .
This isn’t proof that “AI is dumb.” It’s a reminder that not all systems are built the same—and context matters. Carlsen was facing a general-purpose language model, not a chess-trained engine like Stockfish (where he would have zero chances).
My takeaway? In medio stat virtus, as often in life.
AI may fall short in one setup, exceed expectations in another, but the meaningful result often happens in how humans and AI meet, not in who replaces whom.
Have a great weekend!
Giovanni