ClaimsMark Zuckerberg

“I think this year, probably in 2025, we at Meta, as well as the other companies that are basically working on this, are going to have an AI that can effectively be a sort of mid-level engineer that you have at your company that can write code.”

Mark Zuckerberg, Meta CEO, podcast guest 10 January 2025 Transcript of Mark Zuckerberg on Joe Rogan Podcast (The Singju Post)

Prediction AI capabilities

Zuckerberg said he thought that probably in 2025 Meta and other companies working on AI would have an AI that can effectively act as a mid-level engineer writing code.

Unresolvable The public record could not settle it.

Prediction: Statements about events outside the person's control, checked once they fell due: did they come true?

The second judge would not resolve this claim; the verdict shown is the first judge's.

Said for an organisation they lead or work for; it counts on both records.

What would count as true or false

The test is whether, by the end of 2025, Meta had an AI system that could effectively do a mid-level software engineer's job writing code, and other leading AI companies did too. This means working largely on its own on real software-engineering tasks, such as implementing features or fixing bugs across a codebase, rather than only autocompleting code. It should be shown by credible deployment reports (internal or external), demonstrations, or benchmark performance such as solving most tasks on SWE-bench Verified, with reporting that treats it as comparable to a mid-level engineer. True: Meta and at least one other major company had such systems in 2025. Mostly true: such systems clearly existed in 2025 at Meta and elsewhere, but evidence of mid-level-equivalent capability is somewhat weaker for Meta's own system. Mixed: other companies clearly had such agents but Meta did not, or the systems matched a mid-level engineer only on narrow, well-specified tasks. Mostly false: only assistive coding tools, or junior-level agents needing heavy supervision, existed in 2025, or the capability arrived after 2025. False: credible assessments from late 2025 or early 2026 show that no company's AI could act as a mid-level engineer.

Set before the research began.

Evidence (4)

  1. E1 META Q1 2025 Earnings Call Transcript Meta Platforms, Inc. (Investor Relations) · 30 April 2025 Own record

    “So I'd say it's basically still on track for something around a mid-level engineer kind of starting to become possible sometime this year, scaling into next year.”

  2. E2 Claude Sonnet 4.5 Tops SWE-Bench Verified, Extends Coding Focus beyond 30 Hours InfoQ · 11 October 2025

    “On the SWE-bench Verified benchmark, which measures an AI model's ability to solve real-world software issues, Claude Sonnet 4.5 achieved a score of 77.2%, up from 72.7% for Sonnet 4”

  3. E3 2025 Review: How AI for Developers Evolved Neurycode (blog by Yaroslav Dobroskok) · 21 December 2025

    “GitHub Copilot didn't stand still while Cursor took mindshare:”

  4. E4 Mark Zuckerberg Admits Meta's AI Agents Are Behind Schedule ... FourWeekMBA · 6 July 2026

    “In an internal all-hands meeting, Mark Zuckerberg told Meta employees that AI agents have not progressed as quickly as he had hoped.”

Reasoning

The evidence shows a 77.2% SWE-bench Verified score for a non-Meta model [E2] and a later report that Meta's AI agents progressed more slowly than hoped [E4], but no item reports any system as comparable to a mid-level engineer, and E3 does not describe a specific coding agent. The evidence therefore does not settle which outcome under the criteria applies.

Cite this claim

Mark Zuckerberg, “Zuckerberg said he thought that probably in 2025 Meta and other companies working on AI would have an AI that can effectively act as a mid-level engineer writing code.” (said 10 January 2025). Verdict: unresolvable. Reliability API, claim c_c77c9bd5814e, method 0.1.0+3f84d14d, dataset release v1.1.2. https://epistacklabs.com/claims/c_c77c9bd5814e

This claim is in dataset release v1.1.2, with the same verdict.

Reply to this claim

For Mark Zuckerberg, or someone speaking for them: a correction or counter-evidence. Replies are read before anything is shown; an approved reply shows your name, role, date, statement and link. Your email address is never shown and is deleted 30 days after the decision (privacy).

More prediction claims by Mark Zuckerberg

  1. Mixed Prediction 10 January 2025
    “it will also mean that we will dramatically reduce the amount of people whose accounts were taking off for a mistake, which is just a terrible experience.”

    Requiring more confidence for takedowns will dramatically reduce the number of people whose Meta accounts are removed by mistake.

  2. Mostly true Prediction 10 January 2025
    “changing the content filters to have to require higher confidence and precision is actually going to be the thing that reduces the vast majority of the censorship mistakes that we make.”

    Changing Meta's content filters to require higher confidence and precision will reduce the vast majority of Meta's censorship mistakes.

  3. Resolved true Prediction 18 April 2024
    “This is only a matter of time but it's not going to be next year.”

    A 1GW data center will be built eventually, but not in 2025 (next year).

  4. Mostly true Prediction 28 September 2023
    “I think Quest 3 is going to be the first device like this that millions of people are going to get that’s mixed reality.”

    Zuckerberg said he thinks Quest 3 is going to be the first mixed reality device that millions of people get.

All of them

Verdicts are made automatically; how verdicts are made.