Claims that can’t be checked by Gary Marcus, page 18 of 24
Deadlines are judged as of 2026-10-29. About Gary Marcus and the sources covered.
Claims that can’t be checked by Gary Marcus (2353)
Public evidence cannot settle these claims, so they are recorded and attributed, not scored. Most were screened out: both AI providers read the claim on its own and found it a value judgment, rhetoric, hedged, too vague to test, or about something no public evidence could show. The rest were found to be recommendations or too open to test in a full review.
Page 18 of 24: claims 1701 to 1800 of 2353, ordered by the date first said, newest first. To find a claim by its words, use search.
Assessment: Can’t be checkedSaid
Forecast
“Ninth, internal tensions are likely to remain.”
Our reading: Internal tensions at OpenAI are likely to persist in 2024.
No horizon
Assessment: Can’t be checkedSaid
Forecast
“Given how many different kinds of copyrighted works they are drawing on, and from how many vendors (many media outlets, many book publishers, many film studios, etc), and given how commercial their usage is of those sources, the licensing costs may quickly mount.”
Our reading: OpenAI's licensing costs for copyrighted material may quickly mount.
No horizon
Assessment: Can’t be checkedSaid
Forecast
“Second, battles over copyright materials may wind up cutting massively into profits, and may be unavoidable.”
Our reading: Copyright battles may cut massively into OpenAI's profits and may be unavoidable.
No horizon
Assessment: Can’t be checkedSaid
Observation
“In some ways, the most telling example in the earlier newsletter was from Dave Andress (which Vicki Bier shared with us), highlighting the fact that image generators still don’t really understand how the world works:”
Our reading: Image generators still do not really understand how the world works.
No horizon
Assessment: Can’t be checkedSaid
Forecast
“Sometimes GenAI used for source attribution will work, other times it will probably yield outrageous goofs, same as we have seen over and over in other domains.”
Our reading: Generative AI used for source attribution will sometimes work but will probably sometimes produce outrageous errors.
No horizon
Qualifier: used for source attribution
Assessment: Can’t be checkedSaid
Forecast
“Unfortunately, errors of discomprehension may soon be even more worrisome in a new context: war.”
Our reading: AI errors of discomprehension may soon become more worrisome as AI is used in war.
No horizon
Qualifier: in military contexts
Assessment: Can’t be checkedSaid
Rhetoric
“Here is a lovely and instructive set of examples just posted on X by the data scientist Colin Fraser, which serve to remind us that the beauty of LLM’s remains skin deep:”
Our reading: The apparent quality of LLMs is only skin deep.
No horizon
Assessment: Can’t be checkedSaid
Causal claim
“When you have terabytes of images in your training set you can generate really cool images, but that doesn’t guarantee that your system will understand the relation between those images and the world.”
Our reading: Training on terabytes of images enables impressive image generation but does not guarantee the system understands how images relate to the world.
No horizon
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Not sure that’s a great idea, in an era in which AI still can’t quite sort out what’s happening with an elephant on a crowded beach.”
Our reading: Marcus is not sure that hooking AI up to military applications such as drones is a great idea in an era in which AI still can't quite sort out what's happening with an elephant on a crowded beach.
No horizon
Qualifier: not sure
Assessment: Can’t be checkedSaid
Forecast
“Output similarity and how well can it be resolved may well be central in tons of upcoming lawsuits, and as such the outcome could have a huge impact on the future economics of generative AI.”
Our reading: Output similarity may be central in many upcoming lawsuits, and their outcome could greatly affect the economics of generative AI.
No horizon
Source: No, multimodal ChatGPT is not going to “trivially” solve Generative AI's copyright problems
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Gary Marcus still thinks that licensing is the best and only way to go.”
Our reading: Gary Marcus still thinks that licensing is the best and only way to go in dealing with generative AI's copyright problems.
No horizon
Qualifier: stated as Marcus's continuing view ('still thinks')
Source: No, multimodal ChatGPT is not going to “trivially” solve Generative AI's copyright problems
Assessment: Can’t be checkedSaid
Causal claim
“We further speculated that smaller artists might be particularly vulnerable to being infringed; if the system made Mario, everyone might recognize that; if the system ripped off a lesser-known photographer, the user might fail to recognize the original source.”
Our reading: Marcus and Southen speculated that smaller artists might be particularly vulnerable to being infringed by GenAI, because a user might recognize a copy of Mario but might fail to recognize the original source when the system rips off a lesser-known photographer.
No horizon
Qualifier: presented as speculation
Source: No, multimodal ChatGPT is not going to “trivially” solve Generative AI's copyright problems
Assessment: Can’t be checkedSaid
Rhetoric
“The moral of this story, like the moral of so many others I have told, is that almost nothing in AI is truly trivial.”
Our reading: Almost nothing in AI is truly trivial.
No horizon
Source: No, multimodal ChatGPT is not going to “trivially” solve Generative AI's copyright problems
Assessment: Can’t be checkedSaid
Forecast
“(We also speculate that any such classifier would sometimes “false alarm” and misidentify some genuinely new materials as copyrighted.)”
Our reading: Marcus and Southen speculate that any such classifier for detecting copyrighted generated images would sometimes false-alarm and misidentify some genuinely new materials as copyrighted.
No horizon
Qualifier: presented as speculation
Source: No, multimodal ChatGPT is not going to “trivially” solve Generative AI's copyright problems
Assessment: Can’t be checkedSaid
Causal claim
“Fundamentally, Southen and I had argued that our images were evidence not only that some popular GenAI systems had trained on copyrighted images (nobody is really disputing this) but also that users might accidentally find themselves in trouble, because the GenAI systems don’t list sources, and don’t warn users about possible infringment.”
Our reading: Users of popular GenAI systems might accidentally get into legal trouble because the systems do not list sources or warn about possible infringement.
No horizon
Qualifier: because systems don't list sources or warn about infringement
Source: No, multimodal ChatGPT is not going to “trivially” solve Generative AI's copyright problems
Assessment: Can’t be checkedSaid
Conditional forecast
“If the State of California chooses to investigate, one possible outcome could be that the State of California could force OpenAI to dissolve the nonprofit parent, making a distribution, conceivably measured in the billions of dollars, to an external charity (focused in whole or in part on AI).”
Our reading: If the State of California chooses to investigate, one possible outcome could be that it forces OpenAI to dissolve its nonprofit parent and make a distribution, conceivably measured in the billions of dollars, to an external charity focused in whole or in part on AI.
No horizon
Only if: if the State of California chooses to investigate
Assessment: Can’t be checkedSaid
Forecast
“The only thing we can really say for sure is that the litigation is going to be wild.”
Our reading: Copyright litigation over generative AI image and text outputs will be wild.
No horizon
Source: DALL-E’s New Guardrails: Fast, Furious, and Far from Airtight
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“If the current batch of AI companies cannot create AI that performs reasonably well based on public domain sources and whatever they are prepared to pay to license, they should go back to the drawing board—and figure out how to build software that doesn’t have a plagiarism problem—rather than fleecing artists, writers, and other content providers.”
Our reading: If the current batch of AI companies cannot create AI that performs reasonably well based on public domain sources and whatever they are prepared to pay to license, they should go back to the drawing board and figure out how to build software that doesn't have a plagiarism problem, rather than fleecing artists, writers, and other content providers.
No horizon
Only if: if the current batch of AI companies cannot create AI that performs reasonably well based on public domain sources and whatever they are prepared to pay to license
Assessment: Can’t be checkedSaid
Causal claim
“Covering all those claims with the accidental infringing machine they have created would be a nightmare. The only chance they have of becoming the zillionaires they aspire to be is to rewrite the rules.”
Our reading: Indemnifying customers against copyright claims would be a nightmare for OpenAI, and its only chance of great riches is rewriting copyright rules.
No horizon
Assessment: Can’t be checkedSaid
Observation
“Streamers, for example, license everything they stream; there is literally no conceptual reason why OpenAI could not do the same. They just don’t want to.”
Our reading: There is no conceptual reason OpenAI could not license the commercial works it uses, as streamers do; it simply does not want to.
No horizon
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“That’s a flaw in the software, not society’s responsibility.”
Our reading: The plagiarism tendency is a software flaw that society should not be responsible for accommodating.
No horizon
Assessment: Can’t be checkedSaid
Conditional forecast
“if UK or other lawmakers succumb to their nonsense arguments, and rewrite existing rules for the benefit of GenAI, abandoning principles that go back to the time of the printing press, there will be momentary jubilation for many, and OpenAI will (maybe, if the GPU costs and unreliability problems don’t otherwise kill them, two big ifs) get rich.”
Our reading: If lawmakers rewrite copyright rules to benefit GenAI, OpenAI may get rich, provided GPU costs and unreliability problems do not kill it.
No horizon
Only if: if lawmakers rewrite copyright rules for GenAI
Assessment: Can’t be checkedSaid
Observation
“Indeed, as Southen and I pointed out yesterday in the IEEE Spectrum, the tendency towards potential infringement, in current AI looks like maybe it gets worse in bigger models.”
Our reading: The tendency of current AI toward potential infringement may get worse in bigger models.
No horizon
Assessment: Can’t be checkedSaid
Forecast
“Reid Southen and I continued our experiments, and wrote a long paper about what we found, and why they may pose serious problems both for users and developers that are difficult to fix.”
Our reading: What Marcus and Southen found in their experiments with generative AI image systems may pose serious problems, difficult to fix, for both users and developers.
No horizon
Scope: both for users and developers
Source: The potential GenAI copyright infringement plot thickens
Assessment: Can’t be checkedSaid
Forecast
“In all likelihood, the New York Times lawsuit is just the first of many.”
Our reading: The New York Times lawsuit against OpenAI is likely the first of many copyright lawsuits.
No horizon
Source: Things are about to get a lot worse for Generative AI
Assessment: Can’t be checkedSaid
Forecast
“My guess is that none of this can easily be fixed.”
Our reading: The copyright, provenance, and transparency problems of generative AI probably cannot be easily fixed.
No horizon
Source: Things are about to get a lot worse for Generative AI
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“A good system should give the user a manifest of sources; current systems don’t.”
Our reading: Generative AI systems should give users a manifest of sources.
No horizon
Source: Things are about to get a lot worse for Generative AI
Assessment: Can’t be checkedSaid
Observation
“Users may not know when they produce any given image whether they are infringing.”
Our reading: Users of generative AI may not know whether a given image they produce is infringing.
No horizon
Source: Things are about to get a lot worse for Generative AI
Assessment: Can’t be checkedSaid
Observation
“We will publish a full report next week, but it is already clear that what we are finding poses serious challenges for generative AI.”
Our reading: Experimental findings on copyright reproduction pose serious challenges for generative AI.
No horizon
Source: Things are about to get a lot worse for Generative AI
Assessment: Can’t be checkedSaid
Observation
“And OpenAI faces further risks.”
Our reading: OpenAI faces further risks.
No horizon
Source: Things are about to get a lot worse for Generative AI
Assessment: Can’t be checkedSaid
Observation
“It also suggests that Midjourney has been trained on high-resolution copyright images, to which they may or may not have a license.”
Our reading: The evidence suggests Midjourney was trained on high-resolution copyrighted images, which it may or may not have a license for.
No horizon
Qualifier: licensing status unknown
Source: An artist fights back, and Midjourney has embarrassed themselves
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Sure looks like automated, digital plagiarism to me.”
Our reading: To Marcus, Midjourney's output closely resembling copyrighted film stills sure looks like automated, digital plagiarism.
No horizon
Qualifier: sure looks like ... to me
Source: An artist fights back, and Midjourney has embarrassed themselves
Assessment: Can’t be checkedSaid
Causal claim
“It is not inconceivable that AI-generated propaganda could be significantly intensifying the scale and impact of such a campaign.”
Our reading: It is not inconceivable that AI-generated propaganda could be significantly intensifying the scale and impact of a possible foreign (e.g., Russian) active-measures campaign to destabilize the United States through social media posts.
No horizon
Qualifier: not inconceivable that... could (double hedge; raised only as a possibility)
Source: Are social media and coordinated misinformation campaigns fomenting extremism?
Assessment: Can’t be checkedSaid
Causal claim
“The algorithm may be fine when it is directed at dancing bears, but when directed at the daily news, it is likely doing serious damage to our ability to co-exist as humans.”
Our reading: TikTok's recommendation algorithm, when applied to daily news, is likely doing serious damage to people's ability to co-exist.
No horizon
Qualifier: when directed at the daily news
Source: Are social media and coordinated misinformation campaigns fomenting extremism?
Assessment: Can’t be checkedSaid
Observation
“TikTok has no interest in showing anyone a balanced or statistically representative view of the world.”
Our reading: TikTok has no interest in showing users a balanced or statistically representative view of the world.
No horizon
Source: Are social media and coordinated misinformation campaigns fomenting extremism?
Assessment: Can’t be checkedSaid
Recommendation or value judgment
““Going beyond human knowledge” is a bit oversold here.”
Our reading: The claim that FunSearch went 'beyond human knowledge' is a bit oversold.
No horizon
Qualifier: a bit
Source: Sorry, but FunSearch probably isn’t a milestone in scientific discovery
Assessment: Can’t be checkedSaid
Forecast
“As clever as FunSearch is, it’s unlikely to be a major factor in solving cancer or making lightweight batteries.”
Our reading: FunSearch is unlikely to be a major factor in solving cancer or making lightweight batteries.
No horizon
Source: Sorry, but FunSearch probably isn’t a milestone in scientific discovery
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“But then Google DeepMind’s PR team oversold the paper, claiming e.g., that they had “solved a notoriously hard challenge in computing” and ending with this exuberant but immensely speculative set of claims:”
Our reading: Google DeepMind's PR team oversold the FunSearch paper, ending with an exuberant but immensely speculative set of claims.
No horizon
Source: Sorry, but FunSearch probably isn’t a milestone in scientific discovery
Assessment: Can’t be checkedSaid
Observation
“And the solution probably isn’t all that general.”
Our reading: The FunSearch solution probably isn't all that general.
No horizon
Qualifier: probably
Source: Sorry, but FunSearch probably isn’t a milestone in scientific discovery
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“I will give the last word to Filip Piekniewski, who did the right test: try something that isn’t in the vast corpus, that requires real reasoning, and see how it runs…”
Our reading: Filip Piekniewski's test of Gemini, which tried something not in the vast training corpus that requires real reasoning and checked how it ran, was the right test.
No horizon
Scope: Piekniewski's particular test of Gemini
Assessment: Can’t be checkedSaid
Observation
“The video is almost never the true story.”
Our reading: Tech demo videos almost never tell the true story.
No horizon
Scope: tech demo videos in general, judged as a tech industry veteran
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Don’t let the PR fool you.”
Our reading: People should not be fooled by Google's PR about Gemini.
No horizon
Assessment: Can’t be checkedSaid
Observation
“The fact that Google, with all its resources, did NOT blow away GPT-4 could be telling.”
Our reading: The fact that Google, with all its resources, did not blow away GPT-4 could be telling about whether LLMs are close to a plateau.
No horizon
Qualifier: Google, with all its resources, did not blow away GPT-4
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“If the Board was making a step towards bringing them back to their original mission, of AI for human interest rather than for profit, we should be praising them, not burying them.”
Our reading: If the OpenAI board was trying to return OpenAI to its original mission of AI for human interest rather than profit, it should be praised, not condemned.
No horizon
Qualifier: if the board was trying to return OpenAI to its original mission
Source: “Not consistently candid”
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“If he was already working on the chip deal in May at the Senate, that should have been disclosed, too.”
Our reading: If Altman was already working on the Rain chip deal when he testified to the Senate in May 2023, he should have disclosed it.
No horizon
Qualifier: if he was already working on the Rain chip deal in May 2023
Source: “Not consistently candid”
Assessment: Can’t be checkedSaid
Observation
“The more that comes out, the less likely any of these three theories seems to be correct.”
Our reading: The popular theories of the OpenAI board's firing of Altman (D'Angelo conflict of interest, an incompetent board, Toner as a communist) look less and less likely to be correct as more information comes out.
No horizon
Source: “Not consistently candid”
Assessment: Can’t be checkedSaid
Observation
“It may well be worth tens of millions of dollars; even for Altman, that’s not trivial.”
Our reading: Altman's indirect stake in OpenAI through Y Combinator may be worth tens of millions of dollars.
No horizon
Source: “Not consistently candid”
Assessment: Can’t be checkedSaid
Causal claim
“Altman gave up his board seat. If there was really nothing that even looked like legitimate beef against him, I doubt he would have conceded this.”
Our reading: Marcus doubts that Altman would have given up his OpenAI board seat if there had really been nothing against him that even looked like a legitimate grievance.
No horizon
Only if: if there was really nothing that even looked like legitimate beef against him
Source: “Not consistently candid”
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“OpenAI characterizes Altman’s indirect stake as a “small investment”, but OpenAI and Altman should nonetheless disclose the value of that stake, My guess is that it is at least worth many tens of millions, at the new $86B valuation, possibly more, more than many people will earn in a lifetime.”
Our reading: OpenAI and Altman should disclose the value of Altman's indirect stake, which Marcus guesses is worth at least many tens of millions of dollars at the $86B valuation.
No horizon
Qualifier: at the $86B valuation
Source: “Not consistently candid”
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“By any reasonable measure they have strayed, very far, from that mission. That’s primarily on Sam.”
Our reading: By any reasonable measure, OpenAI has strayed very far from its founding mission of safe, broadly beneficial, openly shared AI, and that is primarily Altman's responsibility.
No horizon
Qualifier: by any reasonable measure
Source: “Not consistently candid”
Assessment: Can’t be checkedSaid
Observation
“I submit that we are already part way there.”
Our reading: We are already partway toward a technopolar world where Big Tech companies rather than nations run the world.
No horizon
Source: Critical national security questions should not be decided unilaterally by unelected tech leaders
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“It shouldn’t be entirely up to them to decide. But at present, it is.”
Our reading: Decisions about open-releasing powerful AI should not be entirely up to Meta's leadership, though currently they are.
No horizon
Source: Critical national security questions should not be decided unilaterally by unelected tech leaders
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“With so much at stake, all of this demands careful public discussion. What we should not have is a couple of tech lords sitting around deciding for us.”
Our reading: Whether to open-source powerful AI should be decided through careful public discussion rather than by a few tech leaders.
No horizon
Source: Critical national security questions should not be decided unilaterally by unelected tech leaders
Assessment: Can’t be checkedSaid
Forecast
“Open-source software has been fabulous for the internet, but open-source AI might, for example, be used to design bioweapons.”
Our reading: Open-source AI might be used to design bioweapons.
No horizon
Source: Critical national security questions should not be decided unilaterally by unelected tech leaders
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“The fact that some regulation is bad doesn’t magically mean we should exempt the AI industry; it means we should craft our regulation wisely.”
Our reading: The AI industry should not be exempted from regulation; regulation should be crafted wisely instead.
No horizon
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“I hope that some form of compromise survives, placing genuine requirements on (at least) the most capable foundation models, perhaps empowering a scientific committee to adjust the criterion over time, factoring in size, changes in technology, newly identified risks, the nature of deployment, and so forth.”
Our reading: The EU AI Act compromise should place genuine requirements on at least the most capable foundation models, possibly with a scientific committee adjusting the criterion over time.
No horizon
Assessment: Can’t be checkedSaid
Conditional forecast
“If the EU does not find a compromise, we are all in trouble; the only beneficiaries will be the largest companies who are least in need of help.”
Our reading: If the EU fails to reach a compromise on the AI Act, everyone will be in trouble and only the largest AI companies will benefit.
No horizon
Only if: the EU does not find a compromise
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“We should not allow that to stand.”
Our reading: Society should not allow big tech to exempt generative AI from regulation.
No horizon
Assessment: Can’t be checkedSaid
Forecast
“In the worst case, after five years of negotiation, we could wind up with nothing, leaving the citizens of the world more or less1 entirely on the hook for any negative externalities that arise from generative AI, from misinformation to cybercrime to new bioweapons to bias; runaway AI, if it’s really a thing, would also not be covered.”
Our reading: In the worst case the EU AI Act negotiations could produce nothing, leaving citizens largely unprotected from generative AI's negative externalities.
No horizon
Qualifier: in the worst case
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“The EU needs to cut a deal.”
Our reading: The EU needs to reach a compromise deal on its AI Act.
No horizon
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“So the obvious and correct compromise, which the Spanish government has been trying to push, is a “tiered approach” to try to put the greatest burden on the largest companies.”
Our reading: The EU AI Act should adopt a tiered approach putting the greatest regulatory burden on the largest foundation model companies.
No horizon
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“If you know anyone in the EU government, now is the time to let them know that you don’t want big tech to run the world.”
Our reading: People who know anyone in the EU government should tell them now that they do not want big tech to run the world.
No horizon
Scope: if you know anyone in the EU government
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“AI definitely should not aspire to replicate the human mind in every detail.”
Our reading: AI should not aim to replicate the human mind in every detail.
No horizon
Qualifier: in every detail
Source: AI versus The Clumsy Human Mind
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“The sooner we realize that humans and silicon machines are importantly different, the better.”
Our reading: People should recognize, as soon as possible, that humans and silicon machines are importantly different.
No horizon
Source: AI versus The Clumsy Human Mind
Assessment: Can’t be checkedSaid
Observation
“For now, humans aren’t more intelligent than machines or vice versa, we are just differently abled.”
Our reading: As of November 2023, neither humans nor machines are more intelligent overall; they have different abilities.
No horizon
Qualifier: for now
Source: AI versus The Clumsy Human Mind
Assessment: Can’t be checkedSaid
Causal claim
“It made my day because that’s what I’ve been trying to teach everyone, for a long time: we can probably make machines better if we spend time first thinking about human minds (both their strengths and their weaknesses).”
Our reading: We can probably make machines better if we first spend time thinking about human minds, both their strengths and their weaknesses.
No horizon
Qualifier: probably
Source: AI versus The Clumsy Human Mind
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“That’s what AGI should be: something that solves our problems for us (rather than creating new ones).”
Our reading: AGI should be a reliable, hallucination-free system that solves human problems rather than creating new ones.
No horizon
Source: AI versus The Clumsy Human Mind
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“The point of AI should not be to replicate us, but to supplement us, where we are cognitively frail (calculators are great at that! and spreadsheets! and databases!), and to pitch where we would prefer not to go, helping us, for example, with jobs that robotics call “dull, dirty, and dangerous”.”
Our reading: AI should aim to supplement humans where they are cognitively weak and take on dull, dirty, and dangerous work, not replicate humans.
No horizon
Source: AI versus The Clumsy Human Mind
Assessment: Can’t be checkedSaid
Forecast
“We are probably still many “breakthroughs” away from true level 5 driverless cars.”
Our reading: True level 5 driverless cars are probably still many breakthroughs away.
No horizon
Source: About that OpenAI “breakthrough”
Assessment: Can’t be checkedSaid
Observation
“Although there is some talk of OpenAI already trying to test the new technique out, it wouldn’t be realistic for them to have entirely changed the world in a matter of months.”
Our reading: It is not realistic that OpenAI's Q* technique has already entirely changed the world within a few months.
No horizon
Source: About that OpenAI “breakthrough”
Assessment: Can’t be checkedSaid
Observation
“OpenAI could in fact have a breakthrough that fundamentally changes the world.”
Our reading: It is possible that OpenAI has a breakthrough that fundamentally changes the world.
No horizon
Source: About that OpenAI “breakthrough”
Assessment: Can’t be checkedSaid
Observation
“Part of the joke is that I don’t actually think Bing is powered already by the alleged breakthrough.”
Our reading: Marcus does not think Bing is already powered by OpenAI's alleged Q* breakthrough.
No horizon
Qualifier: I don't actually think
Source: About that OpenAI “breakthrough”
Assessment: Can’t be checkedSaid
Observation
“It’s all still an empirical question; I for one certainly don’t yet know enough about the details of Q* to judge with certainty.”
Our reading: Whether Q* pans out is still an empirical question, and Marcus himself does not yet know enough about the details of Q* to judge it with certainty.
No horizon
Scope: I for one
Source: About that OpenAI “breakthrough”
Assessment: Can’t be checkedSaid
Observation
“At least so far, Q* ain’t one.”
Our reading: At least so far, OpenAI's Q* is not among the problems Marcus is worried about.
No horizon
Qualifier: at least so far
Source: About that OpenAI “breakthrough”
Assessment: Can’t be checkedSaid
Observation
“Me? I am pretty skeptical; I was only half-kidding when I wrote this:”
Our reading: Marcus is pretty skeptical about OpenAI's reported Q* breakthrough.
No horizon
Qualifier: pretty
Source: About that OpenAI “breakthrough”
Assessment: Can’t be checkedSaid
Observation
“GPT’s “lies” are also legendary, and prolific, but not really lies, because real lies intend to deceive, and GPT has no intent.”
Our reading: GPT produces many falsehoods, but they are not really lies because GPT has no intent to deceive.
No horizon
Source: A Tale of Two Liars
Assessment: Can’t be checkedSaid
Observation
“This is inconsistent with all narratives about “cloddish” or self-interested boards, and suggests that something real happened.”
Our reading: The reported concessions by Sam Altman and Greg Brockman (giving up board seats, agreeing to an internal investigation) are inconsistent with narratives that the OpenAI board was cloddish or self-interested.
No horizon
Qualifier: given reports that Sam and Greg gave up board seats and Sam agreed to an internal investigation
Source: Massive scoop
Assessment: Can’t be checkedSaid
Observation
“This is inconsistent with all narratives about “cloddish” or self-interested boards, and suggests that something real happened.”
Our reading: Sam Altman and Greg Brockman giving up their board seats, and Altman agreeing to an internal investigation, are inconsistent with all narratives of a 'cloddish' or self-interested OpenAI board, and suggest that something real happened.
No horizon
Qualifier: based on The Information's report that Sam and Greg gave up their board seats and Sam agreed to an internal investigation
Source: Massive scoop
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“The most important thing for humanity is that OpenAI sticks to its original mission:”
Our reading: The most important thing for humanity is that OpenAI sticks to its original mission.
No horizon
Source: OpenAI is back in business
Assessment: Can’t be checkedSaid
Forecast
“We may never fully know what happened.”
Our reading: The public may never fully learn what happened in the November 2023 OpenAI board crisis.
No horizon
Source: OpenAI is back in business
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Bummed to see no women at all. Hopefully the board will expand in due course.”
Our reading: Marcus is disappointed that OpenAI's new board has no women and hopes the board will expand in due course.
No horizon
Qualifier: hopefully
Source: OpenAI is back in business
Assessment: Can’t be checkedSaid
Forecast
“I am glad to see OpenAI is back in business, and am glad to the continued presence of @AdamDAngelo and addition of two new parties that I expect to be neutral.”
Our reading: Marcus expects the two new members of OpenAI's reconstituted board to be neutral.
No horizon
Source: OpenAI is back in business
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“The fact that the prospectus warned that the whole thing should be viewed “as a donation” doesn’t seem to have much deterred the investors, but it should have.”
Our reading: Investors should have been deterred by OpenAI's prospectus warning that investment be viewed as a donation.
No horizon
Source: Top 5 reasons why OpenAI was probably never really worth $86 billion
Assessment: Can’t be checkedSaid
Forecast
“Of course after all the drama, the company that was reportedly about to invest at a reported $86 billion valuation, Thrive Capital, and its own investors, like CalPERS (the California pension system which is believed to have financed Thrive in part) probably now won’t.”
Our reading: Thrive Capital probably will not now invest in OpenAI at an $86B valuation.
No horizon
Qualifier: after the board drama
Source: Top 5 reasons why OpenAI was probably never really worth $86 billion
Assessment: Can’t be checkedSaid
Observation
“They never really seemed ready to pursue that, but again, there was always a looming reality that they weren’t facing.”
Our reading: OpenAI never seemed ready to move beyond LLMs.
No horizon
Source: Top 5 reasons why OpenAI was probably never really worth $86 billion
Assessment: Can’t be checkedSaid
Rhetoric
“They may have just dodged a bullet.”
Our reading: Thrive Capital and its investors, such as CalPERS, who probably now won't invest at the reported $86 billion valuation, may have just dodged a bullet.
No horizon
Only if: probably now won't invest
Source: Top 5 reasons why OpenAI was probably never really worth $86 billion
Assessment: Can’t be checkedSaid
Forecast
“They may yet get past it, but the tension at the middle would likely remain, always weakening the value of the company as a pure business play.”
Our reading: Even if OpenAI gets past the board drama, the profit-vs-safety tension will likely remain and weaken its value as a business.
No horizon
Qualifier: even if they get past the drama
Source: Top 5 reasons why OpenAI was probably never really worth $86 billion
Assessment: Can’t be checkedSaid
Forecast
“OpenAI might find that new paradigm some day, but someone else (e.g., DeepMind or Anthropic) might well get there first.”
Our reading: Someone other than OpenAI, such as DeepMind or Anthropic, might well find the new paradigm beyond current LLMs before OpenAI does.
No horizon
Source: Top 5 reasons why OpenAI was probably never really worth $86 billion
Assessment: Can’t be checkedSaid
Forecast
“That income is mostly ‘proof of concept” testing, but that income could dry up.”
Our reading: OpenAI's income this year is mostly from proof-of-concept testing, and that income could dry up.
No horizon
Qualifier: mostly
Source: Top 5 reasons why OpenAI was probably never really worth $86 billion
Assessment: Can’t be checkedSaid
Forecast
“The $86B secondary sale to Thrive is presumably unlikely to close.”
Our reading: The secondary share sale of OpenAI to Thrive at an $86B valuation is presumably unlikely to close.
No horizon
Source: Everybody’s out
Assessment: Can’t be checkedSaid
Forecast
“Probably a lot of employees will leave.”
Our reading: A lot of OpenAI employees will probably leave following the breakdown of negotiations to reinstate Sam Altman.
No horizon
Qualifier: following Sam Altman's ouster and breakdown of negotiations
Source: Everybody’s out
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“The schisms on display there highlight the fact that we cannot rely purely on the companies to self-regulate AI, wherein even their own internal governance can be deeply conflicted.”
Our reading: We cannot rely purely on AI companies to self-regulate AI, as the schisms at OpenAI highlight that even their own internal governance can be deeply conflicted.
No horizon
Scope: purely (exclusive reliance on self-regulation)
Source: Open letter to all EU leaders
Assessment: Can’t be checkedSaid
Forecast
“The recent events at OpenAI are likely going to lead to considerable, unpredictable instability.”
Our reading: The November 2023 events at OpenAI are likely to lead to considerable, unpredictable instability.
No horizon
Qualifier: following the recent events at OpenAI (November 2023 leadership crisis)
Source: Open letter to all EU leaders
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Please don't gut the EU AI Act; we need it now more than ever.”
Our reading: EU leaders should not gut (drastically weaken) the EU AI Act, because it is needed now more than ever.
No horizon
Qualifier: we need it now more than ever
Source: Open letter to all EU leaders
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Please don't gut the EU AI Act; we need it now more than ever.”
Our reading: The EU AI Act is needed now more than ever, given the turmoil at OpenAI.
No horizon
Qualifier: in light of the OpenAI turmoil
Source: Open letter to all EU leaders
Assessment: Can’t be checkedSaid
Forecast
“They may well get their way.”
Our reading: OpenAI's investors may well succeed in getting the board to resign and bringing Sam Altman back.
No horizon
Assessment: Can’t be checkedSaid
Conditional forecast
“If Sam doesn’t come back, Mira may rightfully be perturbed, and if the rumor is correct they’ve now undermined her.”
Our reading: If Altman does not return, Mira Murati may rightly be perturbed, and if the rumor is true, the board has undermined her.
No horizon
Only if: if Sam Altman doesn't come back
Source: Firing Sam and then trying to bring him back is a hot mess.
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“From the outside, it’s extremely hard to see how all that adds up to a well-played , well-thought through hand.”
Our reading: From the outside, it is extremely hard to see how the OpenAI board's actions around firing and trying to rehire Sam Altman add up to a well-played, well-thought-through hand.
No horizon
Qualifier: from the outside
Source: Firing Sam and then trying to bring him back is a hot mess.
Assessment: Can’t be checkedSaid
Conditional forecast
“If Sam does come back, they have accomplished nothing (and probably lost Ilya and the $86B secondary sale) and returned to someone they don’t trust, and given him more power, which may inherently undermine safety.”
Our reading: If Altman returns, the board will have accomplished nothing, probably lost Ilya Sutskever and the $86B secondary sale, and given more power to someone it doesn't trust.
No horizon
Only if: if Sam Altman comes back
Source: Firing Sam and then trying to bring him back is a hot mess.
Assessment: Can’t be checkedSaid
Conditional forecast
“If Sam does come back, they have accomplished nothing (and probably lost Ilya and the $86B secondary sale) and returned to someone they don’t trust, and given him more power, which may inherently undermine safety.”
Our reading: If Sam Altman comes back, the board will have given him more power, which may inherently undermine safety.
No horizon
Only if: if Sam does come back
Source: Firing Sam and then trying to bring him back is a hot mess.