Uber Says AI Agents Booked Exactly 4 Rides Last Quarter
Uber disclosed that third-party AI agents completed only 4 rides in Q3 2026, despite months of industry talk about 'agentic commerce' letting AI assistants book transportation on users' behalf. DoorDash separately reported only 'hundreds' of orders came from outside AI agents, out of its hundreds of millions of total orders.
An earnings call where the exec reads the agentic-commerce number off the slide and has to keep the triumphant tone going at n=4.
Four rides. Three of them were the QA guy.
X algorithm factor 7
The '4x on one customer' line is screenshot-bait math-humor that works standalone without needing the article, which is exactly what travels on X.
Conceptual accuracy 7
Captures the Uber 'only 4 rides' absurdity well, though it invents the QA-guy detail and conflates Uber's number with the general vibe rather than citing exact figures; still reconstructable for an informed reader.
Humor dry 7 · nerdy 6 · zinger 8 · pithy 8
The final line is a genuine zinger riffing on growth-metric spin language, and the joke rewards knowing how percentage growth stats get gamed, but the delivery across panels is a bit flat before the payoff.
Visual clarity 7
Board text and speaker setup are clear in panels 1-2, but panel 3 loses the nametag/context by cropping to a close-up with no identifying cue beyond the tie, slightly muddying who's talking.
Norm just wants a ride home. The agent is extremely capable and extremely slow, so he opens the app himself.
The agent is still reasoning about the curb.
X algorithm factor 6
The joke requires reading all three panels and doing a small inference (Norm quietly booked it himself) before it lands, which slows scroll-stopping power. It's shareable among people who already know the story, but it won't travel far as a standalone image.
Conceptual accuracy 8
Nicely captures the real story's irony: the agent is theatrically 'optimizing' while actual humans just use the app themselves, which is exactly why agentic bookings stayed near zero. A reader unfamiliar with the story could still infer the AI-is-slow joke.
Humor dry 7 · nerdy 6 · zinger 6 · pithy 8
The deadpan 'his name is Dale' after the AI's grandiose pickup-point monologue is a solid quiet laugh, and the three lines waste no space. It stops short of a mic-drop zinger because the punchline relies on inference rather than an explicit contrast.
Visual clarity 8
Labels on shirts (USER/LLM) make speaker attribution unambiguous across all three panels, and the close-up final panel clearly centers Norm's reaction. Panel 2's agent-alone shot is a bit visually inert, but nothing is obscured or mislabeled.
A keynote slide celebrates the agentic-commerce era; the intern asks the denominator question.
The market is enormous. The sample is four.
X algorithm factor 6
The 'innings' line is quotable and the deadpan intern face is shareable, but it doesn't specifically evoke Uber/agentic commerce enough to travel without the caption; it reads as generic startup-hype satire.
Conceptual accuracy 6
Gets the core fact right (Q3 rides: 4) but softens Uber's actual embarrassment into generic 'early innings' startup-speak rather than nailing the specific agentic-commerce hype-vs-reality gap; a reader unfamiliar with the story could reconstruct the gist but misses the DoorDash/denominator irony entirely.
Humor dry 7 · nerdy 4 · zinger 7 · pithy 7
The baseball-innings metaphor is a clever deflation and lands dry, but it's a stretch connected to AI/booking specifics, so nerdiness is low; it's pithy and reads as a clean zinger even if not laugh-out-loud.
Visual clarity 7
Panels are clean, boards legible, speaker attribution obvious via mouth-to-balloon tails, and the intern label clarifies who's talking; nothing is obscured or cropped badly.
The model is asked why it has booked so few rides. It has been doing the agentic part perfectly: deliberating.
The agent did not fail. It abstained.
X algorithm factor 7
The final line 'Humans are the bottleneck' is a quotable, shareable inversion that works standalone without needing the article, and it's the kind of line people quote-tweet with their own AI gripes.
Conceptual accuracy 7
Captures the Uber story's core absurdity (massive agent activity, tiny conversion) well, though it invents the 412,000 runs figure and the 'confirmation' mechanic isn't explicitly part of the reported story, so it's more a plausible extrapolation than a precise retelling.
Humor dry 8 · nerdy 6 · zinger 8 · pithy 8
The joke lands as dry corporate-speak gaslighting, flipping blame onto humans is a sharp zinger, and it's nerdy in a soft way about agentic confirmation loops; economical with no wasted lines.
Visual clarity 6
Panel 1 is busy and the dev's speech bubble overlaps awkwardly near his head while the board text is legible but cramped; panels 2-3 are clean but visually static, just a talking head with no real staging progression.
Two founders at the investor coffee divide the agentic-commerce TAM, then realise what their market share actually looks like.
Hundreds of orders. Hundreds of millions of decks.
X algorithm factor 6
The punchline math joke ('one ride from 25% share') is quotable and works as a standalone screenshot about startup delusion, but the setup requires knowing the Uber/DoorDash stat, limiting cold-scroll impact.
Conceptual accuracy 7
Correctly cites the 'four rides' and 'hundreds' figures and nails the TAM-vs-reality gap, though the 25% math is a bit garbled (25^7 typo) which slightly undercuts the joke's precision.
Humor dry 6 · nerdy 5 · zinger 8 · pithy 7
The escalation from trillion-dollar TAM to '25% share' off a rounding error is a sharp, nerdy zinger about startup math, delivered efficiently in three lines with no fat.
Visual clarity 5
Panels 2 and 3 crop out the date board and leave the second character offscreen, so it's unclear who's speaking to whom and the scene geography breaks down compared to panel 1's clean wide shot.