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

Ed Zitron Where's Your Ed At Published

How Does OpenAI Survive?

We recorded 38 claims from this post. 14 were checked against evidence. 24 can’t be checked. 36 more are held. Read the post at its source.

Claims from this post (38)

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: UnresolvedSaid

    “here’s the crux of the matter: generative AI is a product with no mass-market utility - at least on the scale of truly revolutionary movements like the original cloud computing and smartphone booms - and it’s one that costs an eye-watering amount to build and run.”

    What we tested: Generative AI has no mass-market utility on the scale of the cloud or smartphone booms, and it costs an enormous amount to build and run.

  2. Assessment: UnresolvedSaid

    “though I believe OpenAI's continued existence is necessary to keep companies interested/invested in the industry at all.”

    What we tested: OpenAI's continued existence is necessary to keep companies interested and invested in the generative AI industry.

  3. Assessment: SupportedSaid

    “Pre-AGI in this case refers to anything OpenAI has ever developed, as it has yet to develop AGI and has yet to get past the initial "chatbot" stage of its own 5-level system of evaluating artificial intelligence.”

    What we tested: OpenAI has not developed AGI and has not moved past the first, chatbot stage of its own five-level scale.

    What happened: The report places OpenAI just below level 2. Approaching that level does not establish that it had passed the initial chatbot stage. (The Decoder, )

  4. Assessment: Can’t be checkedSaid

    “Sadly, that may be the truth for OpenAI.”

    What we tested: OpenAI may be burning so much cash that it needs a deep-pocketed benefactor to bail it out repeatedly.

  5. Assessment: MixedSaid

    “OpenAI has historically, based on reporting, failed to make its models more efficient, failing to deliver their more efficient "Arrakis" model to Microsoft in late 2023.”

    What we tested: OpenAI has failed to make its models more efficient, including failing to deliver the Arrakis model to Microsoft in late 2023.

    What held: The report describes cancellation of Arrakis after inadequate performance and says Microsoft had hoped it would reduce costs. It supports the specific setback, but does not establish a late-2023 delivery deadline. (The Decoder, )

    What did not: OpenAI explicitly reported achieved efficiency savings before the post. This contradicts a broad claim of failure to improve operating efficiency without inferring savings merely from lower API prices. (OpenAI, )

  6. Assessment: UnresolvedSaid

    “Conversely, OpenAI has devoured no monopoly, and the product category it created — the only one it is really part of — is one entirely subsidized by venture capital and Microsoft.”

    What we tested: OpenAI has displaced no monopoly, and its product category is entirely subsidized by venture capital and Microsoft.

  7. Assessment: Can’t be checkedSaid

    “It is fundamentally not the media’s job to convince the world that OpenAI is a stable company with great things ahead — that’s OpenAI’s job.”

    What we tested: The media should not take on the job of convincing the world that OpenAI is a stable company; that is OpenAI's job.

  8. Assessment: Can’t be checkedSaid

    “Raise more money than anyone ever has before — likely at least $3 billion, but more like $10 billion, and do so soon, likely within the next six months.”

    What we tested: At its current pace, OpenAI would have to raise more money than anyone ever has before, likely at least $3 billion but more like $10 billion, and do so soon, likely within the next six months.

  9. Assessment: UnresolvedSaid

    “Raise either multiple rounds, or the largest funding round ever raised by any company, and then have to keep doing so in perpetuity.”

    What we tested: At its current pace, OpenAI would have to raise either multiple rounds or the largest funding round ever raised by any company, and then keep doing so in perpetuity.

  10. Assessment: Can’t be checkedSaid

    “This isn't impossible, but it feels extremely unlikely that OpenAI would be able to do this for more than a few years.”

    What we tested: It is extremely unlikely OpenAI can keep raising record sums and showing growth to investors for more than a few years.

  11. Assessment: Can’t be checkedSaid

    “As a result, OpenAI's revenue might climb, but it's likely going to climb by reducing the cost of its services rather than its own operating costs.”

    What we tested: Any growth in OpenAI's revenue will likely come from cutting prices rather than cutting its own operating costs.

  12. Assessment: Can’t be checkedSaid

    “Doing so is also likely going to lead to perpetual legal action, especially as 404 Media reports that Runway, a generative video company, likely trained its model on thousands of hours of videos taken from YouTube and pirated sources.”

    What we tested: Acquiring training data for generative models is likely to lead to perpetual legal action.

  13. Assessment: Can’t be checkedSaid

    “And, to be abundantly clear, I am not sure there is enough training data in existence to get these models past the next generation.”

    What we tested: There may not be enough training data in existence to take these models past the next generation.

  14. Assessment: Can’t be checkedSaid

    “But what they are not, at this time, is essential.”

    What we tested: Generative AI models are useful but not essential at present.

  15. Assessment: UnresolvedSaid

    “I see no signs that the transformer-based architecture can do significantly more than it currently does.”

    What we tested: Zitron says he sees no signs that the transformer-based architecture can do significantly more than it currently does.

  16. Assessment: Can’t be checkedSaid

    “It is inherently limited by its architecture, and in turn can only get "better" in a linear fashion.”

    What we tested: Generative AI is inherently limited by its architecture, and as a result it can only get 'better' in a linear fashion.

  17. Assessment: SupportedSaid

    “And OpenAI must grow, because $3.5 billion to $4.5 billion a year in revenue is simply not enough to keep this company going.”

    What we tested: OpenAI's $3.5-4.5 billion annual revenue is not enough to keep the company going.

    What happened: The article reports estimated 2024 operating costs of $8.5 billion against estimated revenue of $3.5–4.5 billion. Both ends of the revenue range fall below the estimated costs, supporting the operating-coverage reading. (The Decoder, )

  18. Assessment: Can’t be checkedSaid

    “it’s tough to make the case that these products will fill the burning shortfall in OpenAI’s balance sheet, and would likely only add to its operational costs”

    What we tested: Products like Sora and SearchGPT are unlikely to fill OpenAI's balance-sheet shortfall and would likely add to its operating costs.

  19. Assessment: UnresolvedSaid

    “If it succeeds — which I add would require potentially entirely new branches of psychology and mathematics to do so, as we humans barely understand our brains — that would be a huge technological achievement that still wouldn't turn things around.”

    What we tested: Even if OpenAI's Strawberry delivered advanced reasoning, it would not turn OpenAI's situation around.

  20. Assessment: Can’t be checkedSaid

    “And, to be clear, it is very, very unlikely this happens.”

    What we tested: It is very unlikely that OpenAI succeeds in building Strawberry with advanced reasoning capabilities.

  21. Assessment: UnresolvedSaid

    “As a result, expectations are higher of what future generations of GPT might do, making it inevitable that the company will disappoint investors and customers.”

    What we tested: OpenAI will inevitably disappoint investors and customers because of inflated expectations for future GPT models.

  22. Assessment: Can’t be checkedSaid

    “While there may be ways to reduce the costs of transformer-based models, the level of cost-reduction would be unprecedented, and likely require entirely new chips, cooling solutions and physical server architecture, none of which OpenAI develops.”

    What we tested: While there may be ways to reduce the costs of transformer-based models, the level of cost reduction would be unprecedented and would likely require entirely new chips, cooling solutions and physical server architecture, none of which OpenAI develops.

  23. Assessment: Can’t be checkedSaid

    “While theoretically Nvidia could produce a much-more-efficient chip, doing so would likely take longer than OpenAI has.”

    What we tested: A much more efficient Nvidia chip would likely take longer to arrive than OpenAI can survive.

  24. Assessment: Can’t be checkedSaid

    “This is a bigger existential threat than a lack of adoption. It means people are using it and not getting enough out of it, which could potentially lead to a significant loss of revenue as the hype cycle dies down.”

    What we tested: Because users are not getting enough out of generative AI, revenue could fall significantly as the hype cycle fades.

  25. Assessment: Can’t be checkedSaid

    “OpenAI is inherently limited by GPT's transformer-based architecture, which does not actually automate things, and as a result may only be able to do "more" and "faster," which does not significantly change the product, at least not in such a way that would make it as valuable as it needs to be.”

    What we tested: OpenAI is inherently limited by GPT's transformer-based architecture, which does not actually automate things, and as a result may only be able to do 'more' and 'faster', which does not significantly change the product, at least not enough to make it as valuable as it needs to be.

  26. Assessment: Can’t be checkedSaid

    “For OpenAI to scale, it would require a massive capital expenditure on multiple levels, chief of them the American power grid (see page 15 of this Goldman Sachs report for a conversation with Microsoft's former VP of energy), which will likely require extensive expansion the likes of which hasn't happened in decades at a time when America is far less apt at infrastructural development.”

    What we tested: For OpenAI to scale, the US power grid would likely need an expansion unseen in decades.

  27. Assessment: Can’t be checkedSaid

    “a frank evaluation of a company that I believe may be teetering on the brink of collapse.”

    What we tested: OpenAI may be teetering on the brink of collapse.

  28. Assessment: Can’t be checkedSaid

    “In the event that it has less than a billion in the bank, OpenAI will likely have to raise $5 billion, and do so in the next three to six months.”

    What we tested: If OpenAI has under $1 billion banked, it will likely have to raise $5 billion within three to six months.

  29. Assessment: UnresolvedSaid

    “It would require a way of expanding what OpenAI can sell to address the entirety of corporate America, which would require use cases that I do not believe generative AI is capable of meeting, like automating chunks of the economy rather than bankrupting freelance designers and copy-editors.”

    What we tested: Generative AI is not capable of the use cases, such as automating chunks of the economy, that OpenAI's survival would require.

  30. Assessment: UnresolvedSaid

    “I am simply saying that for OpenAI to grow to the $10+ billion revenue a year it needs to survive, it would need to replace entire chunks of the labor force.”

    What we tested: OpenAI needs $10+ billion in annual revenue to survive, and to get there it would need to replace whole chunks of the labor force.

  31. Assessment: Can’t be checkedSaid

    “Even if these problems are surmountable, there is simply not enough training data, and even if there were, the cost of processing it will likely vastly outweigh whatever revenue OpenAI makes.”

    What we tested: Even if these problems are surmountable, there is simply not enough training data, and even if there were, the cost of processing it will likely vastly outweigh whatever revenue OpenAI makes.

  32. Assessment: ContradictedSaid

    “Along with $3.5 billion from Saudi Arabia’s Public Investment fund and over $8 billion from Softbank, the latter a secondary market sale, it also raised (equity and debt financing respectively) from Goldman Sachs and Morgan Stanley — two parties I do not believe are going to be willing to subsidize generative AI.”

    What we tested: Goldman Sachs and Morgan Stanley will not be willing to subsidize generative AI.

    What happened: Both named banks participated in OpenAI's executed credit facility, demonstrating willingness to finance the business. The report's commercial interest terms do not establish a concessionary subsidy. (CNBC, )

  33. Assessment: Can’t be checkedSaid

    “Perhaps there’s something I’m missing, but if investors have no path to an IPO, this feels like a game of hot potato, except the loser is left with a big, useless stock.”

    What we tested: Allowing that he may be missing something, Zitron says that if investors have no path to an IPO, investing in OpenAI feels like a game of hot potato in which the loser is left with a big, useless stock.

  34. Assessment: UnresolvedSaid

    “And because OpenAI (and the competition) are so deep in the hole with transformer-based models, I believe they will continue to drive billions into them, burning money on training them using data that may or may not have been legally acquired”

    What we tested: OpenAI and its competitors will keep pouring billions into training transformer models on possibly illegally acquired data.

  35. Assessment: Can’t be checkedSaid

    “It could come up with something new, or find a way to make generative AI much cheaper, but again, that is so very, very unlikely.”

    What we tested: It is very unlikely that OpenAI comes up with something new or makes generative AI much cheaper.

  36. Assessment: Can’t be checkedSaid

    “Doing so multiple times is possible, but very, very unlikely, in the same way that continuing to increase revenue at hundreds of percent each year is possible, but very, very unlikely.”

    What we tested: It is very unlikely that OpenAI will raise huge rounds multiple times or keep growing revenue by hundreds of percent each year.

  37. Assessment: Can’t be checkedSaid

    “I fear OpenAI will compete on price, sending its costs upward, or charge what it needs to to approach break-even, which I don't think it’s willing to do.”

    What we tested: Zitron fears OpenAI will compete on price, sending its costs upward, or else charge what it needs to approach break-even, which he does not think it is willing to do.

  38. Assessment: Can’t be checkedSaid

    “What I am describing here is a deeply unsustainable company that many have pinned their hopes on, and even if I’m wrong, this kind of analysis is necessary when there is a company burning billions of dollars that will likely attempt to absorb billions more dollars of capital to survive.”

    What we tested: OpenAI is a deeply unsustainable company burning billions that will likely try to absorb billions more in capital to survive.

Held from this post (36)

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