Until December (2025), AI was not good enough to do the jobs that companies are now cutting.
Unresolvable The public record could not settle it.
Factual: Statements about the present or the past, checked against the record: did they hold up?
The second judge would not resolve this claim; the verdict shown is the first judge's.
What would count as true or false
Judged as of 2026-03-31. The claim is that before December 2025, AI could not do the jobs companies were cutting as of early 2026, implying AI only became capable around December 2025. True: credible evidence (benchmarks, deployment studies, expert assessments) shows that AI before December 2025 could not perform the main tasks of the roles being cut (for example in software, customer service and back-office work), and that a clear capability jump occurred around December 2025. Mostly true: the general point holds, with a modest error in timing or scope. Mixed: AI was adequate for some of the cut roles before December 2025 and not for others. Mostly false: there is substantial evidence that AI was already performing significant parts of these roles before December 2025 (for example documented replacement of customer-service or coding work in 2023 to 2025), though a capability jump around December 2025 is real. False: the evidence shows AI was widely and adequately doing these jobs well before December 2025, with no notable change then.
Set before the research began.
Evidence (4)
-
E1 Klarna changes its AI tune and again recruits humans for customer service CX Dive · 9 May 2025
“Back in February 2024, Klarna touted impressive figures: The AI assistant had taken over two-thirds of customer service chats — 2.3 million in total — in its first month.”
-
E2 AI in Customer Service (2026): Evidence, Failures and How to Roll It Out AI Magazine · 9 October 2026
“CEO Marc Benioff said in an August 2025 podcast that he had cut support staff from 9,000 heads to about 5,000.”
-
E3 AI in Customer Service (2026): Evidence, Failures and How to Roll It Out AI Magazine · 9 October 2026
“The tool increased productivity, measured as issues resolved per hour, by 15% on average.”
-
E4 AI layoffs are looking more and more like corporate fiction that's masking a darker reality, Oxford Economics suggests Fortune · 7 January 2026
“firms don’t appear to be replacing workers with AI on a significant scale”
Reasoning
The evidence shows one company's statement that its AI assistant handled most customer-service chats in early 2024 [E1]. A Salesforce headcount cut is not attributed to AI [E2], an AI tool assisted workers rather than replacing them [E3], and one report finds firms do not appear to be replacing workers with AI on a significant scale [E4]. No evidence item addresses whether a capability jump occurred around December 2025, so the evidence does not settle the claim under the criteria.
Cite this claim
Marc Andreessen, “Until December (2025), AI was not good enough to do the jobs that companies are now cutting.” (said 31 March 2026). Verdict: unresolvable. Reliability API, claim c_95192d082315, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_95192d082315
This claim is in dataset release v1.1.2, with the same verdict.
Reply to this claim
More factual claims by Marc Andreessen
-
Mixed
“It just can’t have been AI.”
-
Mostly true
“Here's the key: What's happening to the price of AI? It's dropping faster than Moore's Law.”
-
Mostly true
“The revenue growth rate of leading AI companies is definitely faster than any wave I've seen before.”
-
Resolved true
“It's now common for consumer AI to have monthly price tiers of $200 or $300, which I think is very positive.”
Consumer AI products commonly have monthly price tiers of $200 or $300.
-
Resolved true
“But in fact, it's building its own AI model.”
Cursor, though it started by calling large models, is building its own AI model.
Verdicts are made automatically; how verdicts are made.