“And I pointed out that even something as easy as Pac-Man was pretty hard to get right with current tools.”
Our reading:
Current AI coding tools find it hard to correctly build even a simple program like Pac-Man.
Assessment: Unresolved
What happened
Was it true for the period described?
Outcome: Unresolved
We checked, and the evidence so far doesn’t settle it.
Judged on a plain reading. They named no measure, so we used the reading a typical informed reader would take. It is spelled out under “How we judged it”.
Evidence up to . Checked .
Why
The Guardian’s report published March 11, 2025 describes a Grok user encountering missing pellets, broken warp tunnels and problematic ghost behavior [e1]. He reported frustrating debugging and switched to Tetris after an hour [e1]. The same user called the experience “surprisingly easy,” while another reported getting a rough version in seconds that still needed corrections [e2]. These accounts show both rapid initial results and functional problems, but do not settle the general claim about contemporary AI coding tools [e1] [e2].
Caveats
- Both evidence items come from one article about Grok, without a comparison across tools [e1] [e2].
- The publication date is known, but exact trial dates and model versions are not established for March 1–14, 2025 [e1] [e2].
- The accounts do not provide consistent functional checks, reproducible success rates or complete records of corrective work [e1] [e2].
Evidence (2 items)
Context only: The Guardian,
bugs
Why it matters: Supports difficulty.
- Where
- Attempt two
- Publisher
- The Guardian
- Kind of source
- News report
- About events on
- Published
Context only: The Guardian,
easy
Why it matters: Counterpoint.
- Where
- Attempt two
- Publisher
- The Guardian
- Kind of source
- News report
- About events on
- Published
How we judged it
- Kind of claim
- Observation
- Deadline
- None given. We don’t add one.
Conditions
- current tools (their words)
Where we had to interpret
- The ordinary reading concerns material functional difficulty producing a basic playable Pac-Man-style game with contemporary AI coding tools.
- It does not assert impossibility, require an exact arcade reproduction, or imply that every attempt fails.
- Current tools is operationalized as tools available in the first half of March 2025, through the post date.
- A general statement about tools requires more than an isolated failure or an isolated successful demonstration.
Right if: Documented, adequately dated trials across relevant contemporary tools show substantial gameplay errors or considerable corrective intervention under ordinary prompting.
Wrong if: Comparable trials show that those tools reliably produce functionally correct basic implementations with little corrective intervention.
We wrote this test after the outcome was known.
What would settle it
Documented trials from March 1–14, 2025 across relevant tools, recording versions, prompts, attempts, human edits and elapsed effort, with consistent checks of movement, collisions, pellets, enemies and completion.
We found 6 claims in this post. See them all.
Full record
Source and dates
Marcus on AI, Hype, Anthropic’s Dario Amodei, the podcasters who love him — and how the New York Times’ commentary on AI has degenerated into industry cheerleading, paragraph 19.
- Said
- We read the source
- live page, full text, public
- Evidence up to
- Checked
We read this wording on 2026-10-28. The post may have been edited after 2025-03-14.
Where it was said (1)
Said once in the posts we read.
: Marcus on AI, Hype, Anthropic’s Dario Amodei, the podcasters who love him — and how the New York Times’ commentary on AI has degenerated into industry cheerleading
And I pointed out that even something as easy as Pac-Man was pretty hard to get right with current tools.
- Where
- paragraph 19
- Published
- 2025-03-14
- We read it
- , live page
- Access
- full text, public
History
Versions
- Current: Unresolved. Evidence up to 2026-10-28.
What happened to this record
PublishedBy us
Published after both AI reviews and the final check passed.
Held backBy us
The final check failed: a cited source didn’t say what the write-up said.