Pundit ClaimsAI claims by Gary Marcus and Ed Zitron, with the evidence

Ed Zitron Where's Your Ed At Published

Sincerity Wins The War

We recorded 23 claims from this post. 15 were checked against evidence. 8 can’t be checked. 10 more are held. Read the post at its source.

Claims from this post (23)

In the order they appear in the post. Each quotation is the speaker’s words in this post. A label such as “Supported” or “Can’t be checked” is the claim’s current assessment; its page has the evidence and the reasons.

  1. Assessment: Can’t be checkedSaid

    “suffice it to say that it’s highly doubtful that the military is going to rely on an LLM that hallucinates a significant portion of the time.”

    What we tested: It is highly doubtful that the US military will rely on an LLM that hallucinates a significant portion of the time, such as Meta's Llama in the Anduril helmet.

  2. Assessment: SupportedSaid

    “While there are absolutely some jobs being taken by AI, there is, to this point, little or no research that suggests that it’s happening at scale, mostly because Large Language Models don’t really do the things that you need them to do to take someone’s job at scale.”

    What we tested: Although some jobs are being taken by AI, there is little or no research showing AI job displacement is happening at scale.

    What happened: The study covers 11 exposed occupations in Denmark and finds no significant earnings or hours effects. Its literature discussion identifies narrow freelance markets as the closest existing displacement evidence; the event date marks this research entering the contemporary record. (Becker Friedman Institute, University of Chicago, )

  3. Assessment: UnresolvedSaid

    “when those LLMs inevitably shit the bed, those people will be hired back.”

    What we tested: When the LLMs inevitably fail, the workers whom CEOs fired after buying into the AI hype will be hired back.

  4. Assessment: Can’t be checkedSaid

    “These scare tactics exist to do one thing: increase the value of companies like Anthropic, OpenAI, Microsoft, Salesforce, and anybody else outright lying about how “agents” will do our jobs, and to make it easier for the startups making these models to raise funds”

    What we tested: AI job-loss scare tactics exist to do one thing: raise the value of Anthropic, OpenAI, Microsoft, Salesforce and others who are outright lying about agents doing our jobs, and make it easier for AI model startups to raise funds.

  5. Assessment: UnresolvedSaid

    “and then failed because there is no proof that this is actually happening other than that which the media drums up.”

    What we tested: There is no proof, other than what the media drums up, that what Roose's piece describes about AI and entry-level work is actually happening.

  6. Assessment: UnresolvedSaid

    “The “AI is going to eliminate half of white collar jobs” story is one that’s taken hold because it gets clicks and appeals to a fear that everyone, particularly those in the knowledge economy who have long enjoyed protection from automation, has.”

    What we tested: The story that AI will eliminate half of white-collar jobs has taken hold because it gets clicks and appeals to a fear everyone has, especially knowledge-economy workers who have long been protected from automation.

  7. Assessment: ContradictedSaid

    “Christopher Mims of the Wall Street Journal had a story from June 2024 that talked about freelance copy editors and concept artists being replaced by generative AI, but I can find no stories about companies replacing employees.”

    What we tested: Apart from a June 2024 Wall Street Journal story about freelance copy editors and concept artists being replaced by generative AI, he can find no stories about companies replacing employees with it.

    What happened: This published story identifies Dukaan's displaced support employees and Lina's handling of general and account-specific questions. It also reports resolved chats, tickets, and response times; these operational claims are attributed to management, not independently audited. (Human Resources Director, )

  8. Assessment: UnresolvedSaid

    “They can’t, because it isn’t happening at scale, because generative AI does not have the capabilities that people like Dario Amodei and Sam Altman repeatedly act like they do”

    What we tested: Businesses replacing workers with generative AI is not happening at scale.

  9. Assessment: UnresolvedSaid

    “generative AI does not have the capabilities that people like Dario Amodei and Sam Altman repeatedly act like they do”

    What we tested: Generative AI lacks the capabilities Dario Amodei and Sam Altman claim, which is why it is not replacing workers at scale.

  10. Assessment: UnresolvedSaid

    “Can you replace an accountant? No. A doctor? No. A writer? Not if you want good writing.”

    What we tested: Current AI cannot replace an accountant or a doctor, and cannot replace a writer if good writing is wanted.

  11. Assessment: SupportedSaid

    “Generative AI has sped up transcription and translation, which are useful for quick references but can cause genuine legal risk.”

    What we tested: Generative AI has sped up transcription and translation, which are useful for quick reference but can cause genuine legal risk.

    What happened: Whisper took about 8.5 hours against an extrapolated human baseline of 8.8 hours, providing modest support for faster transcription. Its generally usable transcriptions support preliminary-reference usefulness, although this is an inference from accuracy results rather than a direct quick-reference experiment. (International Journal of Social Research Methodology, )

  12. Assessment: UnresolvedSaid

    “Generative AI chatbots are driving people insane by providing them an endlessly-configurable pseudo-conversation too, though that’s less of a “use case” and more of a “text-based video game launched at scale without anybody thinking about what might happen.””

    What we tested: Generative AI chatbots are driving people insane by giving them an endlessly configurable pseudo-conversation.

  13. Assessment: SupportedSaid

    “Software engineers using LLM tools — as I talked about on a recent episode of Better Offline — are finding some advantages, but LLMs are far from a panacea.”

    What we tested: Software engineers get some advantages from LLM tools, but LLMs are far from a panacea.

    What happened: The March survey of 3,500 developers and managers found widespread reported time savings from generative AI alongside substantial time lost to organizational inefficiencies. Both findings directly fit benefits coexisting with important unresolved problems. (Atlassian, )

  14. Assessment: Can’t be checkedSaid

    “Let’s be real: none of this is transformative.”

    What we tested: None of the generative AI uses he just listed (transcription, translation, video editing, summarising, coding help, chatbots) is transformative.

  15. Assessment: SupportedSaid

    “An aside about software engineering: I actually believe LLMs have some value here.”

    What we tested: LLMs have some value in software engineering.

    What happened: Pichai identifies the measure immediately beforehand as Google's engineering velocity improvement from AI, rather than the share of generated code. This contemporaneous company estimate supports the modest claim of some practical value. (Lex Fridman Podcast; statement by Google's CEO, )

  16. Assessment: Can’t be checkedSaid

    “while this might affect entry-level software engineers at some point”

    What we tested: LLM coding tools might affect entry-level software engineering jobs at some point.

  17. Assessment: UnresolvedSaid

    “there’s yet to be proof it’s actually happening, or that AI’s taking these jobs and not, say, outsourcing.”

    What we tested: There is not yet proof that AI is taking entry-level software engineering jobs, as opposed to, for example, outsourcing.

  18. Assessment: UnresolvedSaid

    “The word “powerful” is a marketing term that the media has adopted to describe something it doesn’t understand, along with the word “agent,” which means “autonomous AI that can do things for you” but is used, at this point, to describe any Large Language Model doing anything.”

    What we tested: 'Powerful' and 'agent' are marketing terms the media has adopted to describe something it doesn't understand, and 'agent', which means autonomous AI that can do things for you, is now used for any LLM doing anything.

  19. Assessment: Can’t be checkedSaid

    “But the intention is to frame these models as “powerful” and to use the term “agents” to make this technology seem bigger than it is, and the people that control those terms are the AI companies themselves.”

    What we tested: The intention behind calling these models 'powerful' and using the term 'agents' is to make the technology seem bigger than it is, and the AI companies themselves control those terms.

  20. Assessment: Can’t be checkedSaid

    “It’s a disgrace, and I believe it either creates distrust in the media or drives people insane as they look at reality - where generative AI doesn’t really seem to be doing much - and get told something entirely different by the media.”

    What we tested: In reality generative AI does not seem to be doing much, despite what the media says.

  21. Assessment: Can’t be checkedSaid

    “What’s crazy to me is that many of the problems with the AI bubble — as with the metaverse, as with the return to office, as with inflation and price gouging — are obvious if you actually use the things or participate in reality, but such things do not always fit with the editorial message.”

    What we tested: Many of the problems with the AI bubble are obvious if you actually use the products or participate in reality, but they do not always fit the editorial message.

  22. Assessment: Can’t be checkedSaid

    “By not saying how unlikely it is — again, without even saying the word “unlikely,” but allowing the audience to come to that conclusion by having the whole story — you give the audience the truth.”

    What we tested: It is unlikely that acquiring Jony Ive's hardware startup will add $1 trillion in market value to OpenAI, as Sam Altman claimed.

  23. Assessment: UnresolvedSaid

    “The products they use every day are increasingly-broken, pumped full of generative AI features that literally get in the way of what they’re trying to do, which already was made more difficult because companies like Meta and Google intentionally make their products harder to use as a means of making more money.”

    What we tested: Everyday products are increasingly broken and stuffed with generative AI features that get in the way of what users are trying to do.

Held from this post (10)

These records did not pass every check, so nothing about them is published. Held is not a finding.