Unresolved claims, page 17 of 20

Claims by Gary Marcus and Ed Zitron. Deadlines are judged as of 2026-10-29.

Unresolved claims (1958)

The record was checked, and the evidence available by the cutoff does not settle it. Forecasts whose deadline has not arrived are listed apart, as not yet due.

Page 17 of 20: claims 1601 to 1700 of 1958, ordered by the date first said, newest first. To find a claim by its words, use search.

Assessment: UnresolvedSaid 2024-02-19
Ed Zitron
“What these stories don't seem to discuss are whether these companies are making a profit, likely because generative AI is a deeply unprofitable product, demanding massive amounts of cloud computing power”
Assessment: UnresolvedSaid 2024-02-19
Ed Zitron
“Tech's largest cash cow since the cloud computing boom of the 2000s is based on a technology that is impossibly unreliable, a technology with a potent inverted Midas touch that burns far more money than it makes.”
Assessment: UnresolvedSaid 2024-02-18
Gary Marcus
“We are about to have a whole generation of children educated by fake videos that are completely plausible to naive audiences yet biologically (or physically or culturally, etc) incorrect.”
Assessment: UnresolvedSaid 2024-02-17
Gary Marcus
“What’s missing is any global comprehension by the system.”
Assessment: UnresolvedSaid 2024-02-17
Gary Marcus
“Rather, Sora is failing to apprehend a cultural regularity of the world, despite ample evidence.”
Assessment: UnresolvedSaid 2024-02-17
Gary Marcus
“But it’s drawing the wrong generalization.”
Assessment: UnresolvedSaid 2024-02-17
Gary Marcus
“Sora fundamentally revolves around trying to take pixels and reconstruct what pixels might be nearby in space and time. The forest is never perceived apart from the trees, and the chessboard is never perceived apart from its constituent squares.”
Assessment: UnresolvedSaid 2024-02-16
Gary Marcus
“Because the glitches don’t stem from the data, they stem from a flaw in how the system reconstructs reality.”
Assessment: UnresolvedSaid 2024-02-16
Gary Marcus
“More data won’t solve that problem.”
Assessment: UnresolvedSaid 2024-02-16
Gary Marcus
“And importantly, I predict that many will be hard to remedy.”
Assessment: UnresolvedSaid 2024-02-15
Gary Marcus
“If the system really is being “programmed in English” it is not always doing what is being asked to do:”
Assessment: UnresolvedSaid 2024-02-15
Gary Marcus
“Much of what Reid Southen, Katie Conrad, and I recently described is in flagrant violation of those directives.”
Assessment: UnresolvedSaid 2024-02-13
Gary Marcus
“As Ernie Davis, Scott Aaronson, and I showed, a week or so after the release, it’s easily fooled, especially with complex sentences and words like “similar” that would require a deep understanding of language to be interpreted correctly”
Assessment: UnresolvedSaid 2024-02-13
Gary Marcus
“The lesson I drew from from that, back then, in May 2022, was the problem lies not in Dall-E’s drawing ability, but in its ability to cope with noncanonical prompts.”
Assessment: UnresolvedSaid 2024-02-13
Gary Marcus
“One winner out of three to be exact.”
Assessment: UnresolvedSaid 2024-02-13
Gary Marcus
“Again, here is my first and only try. Not a single horse riding an astronaut:”
Assessment: UnresolvedSaid 2024-02-13
Gary Marcus
“In other words, the problem was with its language understanding, rather than with illustration per se. The model was perfectly capable of drawing a horse riding an astronaut, it just didn’t know that the phrase horse rides astronaut was a request for a horse riding an astronaut.”
Assessment: UnresolvedSaid 2024-02-13
Gary Marcus
“I think we can safely reach the same conclusion as before. Gemini can draw the desired image. But it’s still like pulling teeth to get there.”
Assessment: UnresolvedSaid 2024-02-13
Gary Marcus
“In our own brief experiments with GPT-3 (OpenAI has refused us proper scientific access for over a year) we found cases like the following, which reflects a complete failure to understand human biology.”
Assessment: UnresolvedSaid 2024-02-13
Gary Marcus
“That’s why statistically improbable requests like horses riding astronauts have always been challenging for generative image systems.”
Assessment: UnresolvedSaid 2024-02-13
Gary Marcus
“Once again statistical frequency (most guitarists play righthanded) won out over linguistic understanding, to the consternation of Hendrix fans everywhere.”
Assessment: UnresolvedSaid 2024-02-10
Gary Marcus
“Already there is pressure from GenAI to keep coal plants on.”
Assessment: UnresolvedSaid 2024-02-10
Gary Marcus
“Negative externalities, essentially none of which OpenAI is absorbing, like the cost of dis- and misinformation, cybercrimes, and the latest, fake books, a scam that has recently hurt at least three authors I know over the last two months.”
Assessment: UnresolvedSaid 2024-02-09
Gary Marcus
“They are already a target for attack, and as long as they are successful they will continue to be; if they have direct access to user devices, the risks of them getting hacked will increase, and the spoils of war for the attackers will increase, since it enables essentially infinite mischief (e.g., draining bank balances, stealing credit card numbers, placing fraudulent transactions, extortion based on private photos, holding devices for ransom, and so on).”
Assessment: UnresolvedSaid 2024-02-09
Gary Marcus
“Chatbots are struggling to make the money people envision, so he is doubling down to (a) agents, and (b) a multitrillion-dollar chips-and-energy infrastructure play.”
Assessment: UnresolvedSaid 2024-02-08
Gary Marcus
“GenAI systems do effectively store things. (In this case, canonical video game controller images.)”
Assessment: UnresolvedSaid 2024-02-08
Gary Marcus
“Noncanonical situations continue to remain difficult for generative AI systems, precisely because frequent statistics are a mediocre substitute for deep understanding.”
Assessment: UnresolvedSaid 2024-02-08
Gary Marcus
“They (often) don’t come up with a novel independent design that would meet the requirements.”
Assessment: UnresolvedSaid 2024-02-08
Gary Marcus
“Scale isn’t everything, and hasn’t and won’t be enough to solve any of the problems I just listed above.”
Assessment: UnresolvedSaid 2024-02-08
Gary Marcus
“The question is not whether a clever user with a lot of experience and time to kill can induce a mission to do something correctly, but whether the systems reliably deliver what they are asked for. They don’t.”
Assessment: UnresolvedSaid 2024-02-07
Gary Marcus
“But anyone who fears that GPT-5 is going to convey an immense advantage in military strategy has got to be kidding.”
Assessment: UnresolvedSaid 2024-02-07
Gary Marcus
“Even with a lot of data, GPT-4 is playing chess at around the level where (top) chess programs were in 1976, not even close to current state of the art.”
Assessment: UnresolvedSaid 2024-02-05
Gary Marcus
““Given vast datasets, LLMs approximate well, but their understanding is at best superficial. That’s *why* they are unreliable, and unstable, hallucinate, are constitutionally unable to fact check, etc.””
Assessment: UnresolvedSaid 2024-02-05
Gary Marcus
“OpenAI will no doubt by now have filtered this specific example so that it can no longer be replicated, now that it has appeared in a lawsuit.”
Assessment: UnresolvedSaid 2024-02-05
Gary Marcus
“Because of the immense amounts of data they are now trained on, and because of the immense number of parameters on which they are trained, that reconstruction can in fact sometimes come very close to memorization.”
Assessment: UnresolvedSaid 2024-02-05
Gary Marcus
“LLMs do (sometimes) regurgitate large chunks of effectively stored text, and they don’t understand anything at a deep level.”
Assessment: UnresolvedSaid 2024-02-04
Gary Marcus
“At present, we do not know.”
Assessment: UnresolvedSaid 2024-02-04
Gary Marcus
“It is shocking that neither OpenAI nor the media noted this (they also did not discuss in any detail one other paper that reached a somewhat similar conclusion, by Kevin Esvelt’s lab).”
Assessment: UnresolvedSaid 2024-02-01
Gary Marcus
“Not every image looks like this, but it’s pretty clear after a few minutes experimentation from a friend that (a) Bard makes fantastic images but that (b) it is drawing from copyrighted sources and (c) as capable as the other systems Reid Southen and I looked at in returning derivative artwork that appears to infringe without attribution.”
Assessment: UnresolvedSaid 2024-01-30
Ed Zitron
“We are beset with spam calls, spam texts, spam messages on social networks, notifications from apps we forgot we installed, search results that are stuffed with results built specifically to manipulate search engines, emails from companies we’ve bought one thing from, popular home pages full of AI-spun misinformation and algorithms that try and hurt and manipulate us.”
Assessment: UnresolvedSaid 2024-01-29
Gary Marcus
“By my count (helped considerably by my sharp-eyed followers on X), there are at least ten ways, many subtle, in which the image neither matches the prompt nor what we would expect from the world.”
Assessment: UnresolvedSaid 2024-01-29
Gary Marcus
“Hallucinations come from systems exploding data into tiny bits, and then reconstructing them without having a mechanism like a fact-checker to assess the coherence of those reconstructed bits.”
Assessment: UnresolvedSaid 2024-01-29
Gary Marcus
“Near-plagiarism comes from reconstructing those bits in statistically probable ways, without having a mechanism like an originality-assessor to investigate the novelty of the reconstruction.”
Assessment: UnresolvedSaid 2024-01-27
Gary Marcus
“Third, at least so far OpenAI lacks both profits and a moat.”
Assessment: UnresolvedSaid 2024-01-27
Gary Marcus
“Many companies are still in wait and see mode, trying the software but not putting into full-time production.”
Assessment: UnresolvedSaid 2024-01-25
Gary Marcus
“Such errors clearly aren’t disappearing any time soon.”
Assessment: UnresolvedSaid 2024-01-25
Gary Marcus
“Examples like these, in which ChatGPT confidently generates an image that does not actually meet requirements, are nicely complement the visual hallucinations that Ernest Davis and I reported here in October, in which visual ChatGPT misinterpreted various images:”
Assessment: UnresolvedSaid 2024-01-24
Gary Marcus
“It might be trivial to do it for the most popular characters, presented in the most canonical ways, but getting it right for all copyright materials is an entirely different matter.”
Assessment: UnresolvedSaid 2024-01-24
Gary Marcus
“Reid spent about 45 minutes looking into this earlier today (why not more? see below) with images generated by Midjourney, and it quickly became apparent that using ChatGPT as intern to trace down sources is, well, pretty intermittent.”
Assessment: UnresolvedSaid 2024-01-16
Ed Zitron
“The podcast has also struggled to retain its initial momentum, with viewership tumbling from 160,000 views on the first episode to around 40,000 an episode for the last 9 months, with the exception of the George Carlin special which sits at 55,000 views at the time of writing.”
Assessment: UnresolvedSaid 2024-01-10
Gary Marcus
“Notice, too, that Katie asked Bing “are they original?”, and Bing, untruthfully, but with a cozy emoticon, answered in the affirmative.”
Assessment: UnresolvedSaid 2024-01-10
Gary Marcus
“Fundamentally, virtually any guardrail has to thread a needle between the Scylla of being too restrictive and Charybdis of being too permissive. None thus far have done this effectively.”
Assessment: UnresolvedSaid 2024-01-10
Gary Marcus
“Any one example can be filtered out, but there are too many examples, and current systems are too erratic to reliably get things right.”
Assessment: UnresolvedSaid 2024-01-10
Gary Marcus
“Guardrail builders generally wind up playing whack-a-mole.”
Assessment: UnresolvedSaid 2024-01-10
Gary Marcus
“Other prompts still elicited apparent copyright infringements, where new filters did not suffice.”
Assessment: UnresolvedSaid 2024-01-10
Gary Marcus
“Since that would violate the Berne Convention and infuriate every “content provider” (from individual artists and writers to major studios and publishers) ever, we hope and expect that this exemption will not be granted.”
Assessment: UnresolvedSaid 2024-01-08
Gary Marcus
“It’s likewise a flaw that that current AI cannot learn without a tendency to inadvertently regurgitate big bits of text and images that look uncannily like their sources, but not a logical necessity.”
Assessment: UnresolvedSaid 2024-01-08
Gary Marcus
“LLMs may eventually be replaced by better technologies, but for now, they aren’t going anywhere, even if some of the commercial purveyors go out of business.”
Assessment: UnresolvedSaid 2024-01-08
Gary Marcus
“But in the end, nobody will create good, fresh, new content anymore, and the internet will start eating its own tail. We will will all suffer in the end, fed a mediocre stew of regurgitate.”
Assessment: UnresolvedSaid 2024-01-08
Gary Marcus
“OpenAI’s lobbying campaign, simply put, is based on a false dichotomy (give everything to us free or we will die)—and also a threat: either we get to use all the existing IP we want for free, or you won’t get to generative AI anymore.”
Assessment: UnresolvedSaid 2023-12-29
Gary Marcus
“Generative AI systems are fully capable of producing materials that infringe on copyright.”
Assessment: UnresolvedSaid 2023-12-29
Gary Marcus
“And worse, infringement can happen even the user isn’t looking to infringe and doesn’t mention any character or film by name:”
Assessment: UnresolvedSaid 2023-12-29
Gary Marcus
“Unless and until somebody can invent a new architecture that can reliably track provenance of generative text and/or generative images, infringement – often not at the request of the user — will continue.”
Assessment: UnresolvedSaid 2023-12-29
Gary Marcus
“OpenAI’s image software (which we accessed through Bing) is perfectly capable of verbatim and near-verbatim repetition of sources as well.”
Assessment: UnresolvedSaid 2023-12-29
Gary Marcus
“They do not inform users when they do so.”
Assessment: UnresolvedSaid 2023-12-23
Gary Marcus
“Before the day was done, Midjourney retaliated - by revoking Southen’s privileges and wiping his history.”
Assessment: UnresolvedSaid 2023-12-23
Gary Marcus
“The output was not marked as being directly premised on copyrighted work”
Assessment: UnresolvedSaid 2023-12-12
Gary Marcus
“But the big tech executives that would need to deal with it aren’t happy, and are doing their best after the fact to undermine a hard-won compromise that involved years of negotiations and multiple all-nighters.”
Assessment: UnresolvedSaid 2023-12-09
Gary Marcus
“What I proposed then, as a replacement to the Turing Test, was a Comprehension Challenge; even now, no software would be able to meet that challenge.”
Assessment: UnresolvedSaid 2023-12-09
Gary Marcus
“Dubious Google Gemini videos aside, nobody is close to passing that yet.”
Assessment: UnresolvedSaid 2023-12-07
Gary Marcus
“What we are going to see, in the fullness of time, I promise you, is that Gemini is more or less in the same ball park as GPT-4, handy for a bunch of things, but untethered in reality, still with dicey, unpredictable reasoning, and a very limited understanding of the world.”
Assessment: UnresolvedSaid 2023-12-07
Gary Marcus
“Regular readers will recall how that turned out: the Rubik’s cube was instrumented with sensors, and not straight off the shelf, and what worked for the Rubik’s cube in the demo video didn’t turn out have a whole lot of application in the real world.”
Assessment: UnresolvedSaid 2023-12-06
Gary Marcus
“From a commercial standpoint GPT-4 is no longer unique.”
Assessment: UnresolvedSaid 2023-12-05
Ed Zitron
“Already, despite only being available to a handful of users, Grok has fabricated a story about a British MP being expelled from Parliament, presumably confusing the ouster of disgraced New York congressman George Santos with a separate and unrelated story from across the Atlantic.”
Assessment: UnresolvedSaid 2023-12-05
Ed Zitron
“Musk rushed a Large Language Model to market in a few months, one where bigotry is a feature not a bug, and has potentially trained it on the garbled musings of people that he has incentivized to fill its data source with nonsense.”
Assessment: UnresolvedSaid 2023-12-05
Ed Zitron
“Musk is already using Grok as a marketing vehicle for X Premium+’s $16-a-month subscription, and if he actually offers it to every subscriber, there’s no way in hell that this is a profitable transaction.”
Assessment: UnresolvedSaid 2023-12-05
Ed Zitron
“If Musk launches Grok to every Premium+ subscriber, it creates an inverse revenue model where the site becomes less profitable as its products become more popular.”
Assessment: UnresolvedSaid 2023-11-28
Gary Marcus
“So far, Big Tech rather than the US government is calling the shots.”
Assessment: UnresolvedSaid 2023-11-28
Gary Marcus
“But virtually all existing laws were written before AI was really on the scene, and so very few really envision how AI changes things.”
Assessment: UnresolvedSaid 2023-11-28
Gary Marcus
“Big tech wants the world to revolve around its new plaything, Generative AI (combined total revenue so far a few billion dollars, just a tiny tiny fraction of the economy), and to exempt it from any regulation.”
Assessment: UnresolvedSaid 2023-11-27
Gary Marcus
“Actually, there was one clear overall winner: Tegmark. As much as Ernie Davis and I utterly disagreed with Tegmark’s more technical map of the world/LLM world models paper a few weeks ago, pretty much every word he said this week about regulation was true.”
Assessment: UnresolvedSaid 2023-11-27
Gary Marcus
“Bing,powered by Dall-E 3, was an even bigger loser, drawing “Hinton vs LeCun vs Ng vs Tegmark, boxing” with four boxers not boxing, not one of them looking remotely like any of our heroes:”
Assessment: UnresolvedSaid 2023-11-25
Gary Marcus
“Eventually, we will be able to build machines that help us reason critically, overcome human bias, and objectively, the way that we have calculators that spit out just the facts, without confabulation.”
Assessment: UnresolvedSaid 2023-11-23
Gary Marcus
“But “breakthroughs” rarely turn to be general to live up to initial rosy expectations. Often advances work in some contexts, not otherwise.”
Assessment: UnresolvedSaid 2023-11-22
Gary Marcus
“It autocompletes sentences, some of those completions are true given that the training data is so vast, and some aren’t, given that the training data aren’t infinite.”
Assessment: UnresolvedSaid 2023-11-22
Gary Marcus
“The real value was always the people, not the IP, not the data, not the customer list, not the infrastructure.”
Assessment: UnresolvedSaid 2023-11-22
Gary Marcus
“The whole nonprofit-hybrid model, introduced in 2019 (the company was originally pure non-profit) was always a ticking time bomb, because the people trying to make money were never going to have a perfect relationship with those in the nonprofit that were charged with keeping the world safe.”
Assessment: UnresolvedSaid 2023-11-22
Gary Marcus
“Large language models alone were never going to solve the alignment problem.”
Assessment: UnresolvedSaid 2023-11-20
Ed Zitron
“OpenAI was already in a crisis before Altman was fired.”
Assessment: UnresolvedSaid 2023-11-20
Ed Zitron
“With AI development an incredibly costly business, OpenAI can’t afford a tightening of the capital floodgates.”
Assessment: UnresolvedSaid 2023-11-20
Ed Zitron
“Shear doesn’t have much of an AI background, and has never faced the regulatory and legal challenges that threaten OpenAI and the wider AI sector.”
Assessment: UnresolvedSaid 2023-11-05
Gary Marcus
“Essentially NONE of the comparisons we have heard between humans and machines have looked at cars in which humans were not playing some role in the driving, either in the car or remotely.”
Assessment: UnresolvedSaid 2023-11-05
Gary Marcus
“Uber and Lyft can stop worrying about being disintermediated by machines; they will still need human drivers for quite some time.”
Assessment: UnresolvedSaid 2023-11-05
Gary Marcus
“High-reliability is not their forte. and nor are outliers. They won’t take humans out of the loop.”
Assessment: UnresolvedSaid 2023-11-04
Gary Marcus
“Certainly no commercial product could ever work at a profit if you needed remote operators anything like that often.”
Assessment: UnresolvedSaid 2023-10-30
Gary Marcus
“How effective this actually becomes depends a lot on the exact wording, and how things are enforced, and how much of it is binding as opposed to merely voluntary.”
Assessment: UnresolvedSaid 2023-10-30
Gary Marcus
“Companies will poke for loopholes and surely find them.”
Assessment: UnresolvedSaid 2023-10-27
Gary Marcus
“We will eventually see search engines that return neatly written paragraphs, and eventually see AI agents that can organize lives. Or at least our grandchildren or their children will.”
Assessment: UnresolvedSaid 2023-10-27
Gary Marcus
“To get there, though, what we need is a much smarter AI, a kind that can reason, and not hallucinate, grounded in reality and not just corpus statistics— which require fundamental research of a sort that is currently getting starved out in the LLM euphoria.”
Assessment: UnresolvedSaid 2023-10-26
Gary Marcus
“Current AI has no way to reliably anticipate consequences of actions that it might recommend, Asimov’s First Law of Robotics called for robots to do no harm, but current AI has no reliable way to anticipate harm.”