I heard a rumor (on instagram, so YMMV) that the professor who was closest to solving this problem had only weeks ago used Codex, which had slurped up all his notes on the subject. Now OpenAI's agents solve the problem. If it's true that seems like quite a coincidence.
See, e.g., Barak Ravid regularly reporting in Axios the impending ceasefire negotiation progress in the Iran War, which largely have failed to come to pass.
I won't take a side in things, but OpenAI stated the model they used here started training August 28th. Note that "training" here might mean "post-training with RLHF an Astra base model" or something. but training had only started a little over a week earlier.
IMO It’s not about being trained on the data, it’s more like what do the agents have access to during inference? Can they grep customer transcripts/logs?
Has it not been the usual process to snapshot a model to use for inference while continuing to run the training process? I guess you can’t add to the training corpus once you begin? Just trying to make sense of whether training begins or ends as rigidly as you suggest.
No, it's quite well defined, and it's not called a 'memory subsystem'. There is a training process (as in traditional Machine Learning training) that occurs with data available at the specific point in time the training starts (or ends), this is called the cutoff date.
After this point there can be other kinds of training, the weights can be shifted, the internal CoT prompts can be changed, routing in MoE can change, but the Foundational Model that was trained on a corpus is the same model trained in the same corpus.
User data can be used at any of these stages theoretically of course, but by the nature of training and from the dates of the events, (a new Foundational Model being released), it would look as if the user data of the professor was used in the training of the foundational model, which is something that OAI does every couple of months for a big release, and incorporates the new text from their text scraping efforts, including new books ingested, new internet text scraped, deals with third party platforms, and data from their own users (not conjectured, read the ToS, users allow this.)
There have been several threads and developments on this over the past few days, including statements from the primary subjects involved. Third-hand instagram comments are not really the best source to be bringing in.
Everybody comes into information in different ways. There were no comments here about this specific aspect of the story - which is definitely interesting!
Location: Nelson, New Zealand
Remote: Yes
Willing to relocate: No
Technologies: Rust, graphics via `wgpu`, WebGPU, frontend, backend, retail, AI inference on edge, Zcash crypto experience (with no desire to continue in crypto)
Resume: https://zyghost.com/resume.html
Email: efsubenovex@gmail.com
GitHub: https://github.com/schell
I'm a generalist that has worked in many domains, with 20+ years of experience. I'm looking to get out of crypto. Senior or Staff level Engineering (IC) is a good fit.
Came here to say exactly this. If you’re not writing the code (or especially reviewing it) then we need stronger type systems and more checks and fewer legal programs. Might as well move all the way to Idris or some not-yet-invented language that humans would find very restrictive.
This should tell you how unserious these "companies" are when they consider themselves the grand harbinger of the one and true "AGI" because their frontier models were able to solve (with great assistance) some sort of conjecture that was never taken seriously in the mathematical world but somehow is significant now because ???
I was wondering where this goal post would shift to next. So now instead of AI being incapable of doing something not in its training data it needs too much external help doing what no human could do, and what it came up with isn't even important anyways.
The companies can suck and their product can do amazing things at the same time.
It is interesting to see CEOs of big tech companies are so convinced they are always right that they start sidestepping any internal review of communication. Often to the detriment of the company.
“Nothing in my training data covers this, but if I had to guess, OpenAI may have released information about GPU support in Mac OS X early, pissing off Steve Jobs.”
Which, somehow, confused me even further. I get that it's a stylistic choice to make them easy to read or whatever, but my first thought was "wait, they just posted screenshots with the names censored like a bad Reddit post?"
This is pretty wild. Without context, it's barely coherent, but I guess they didn't want to say "Apple is suing us for X, Y, Z" because that makes them look bad. But there's just no upside to publishing this:
If you don't know what's going on, this blog post is nonsensical and makes OpenAI look weird.
If you do know what's going on, you'll immediately notice they don't address 80% of the stuff Apple is accusing them of, which makes OpenAI look dishonest.
It just seems incredibly unprofessional and ill-considered.
It's always suspect when one side tries to argue their case in the court of public opinion.
Frankly, it's a big red flag for me and immediately casts suspicion on said party. I mean, do you think you're going to lose in court and so you're trying to get out ahead of that?
No. This is a fucking legal case. It will be decided in the courtroom, not on the internet.
Posting this is insane in the sense of "What will posting this possibly improve for OpenAI?"
It's not going to change the court case. It's not going to impress professional adults.
I guess... teenage reddit users with too much time on their hands who are fans of OpenAI? Though struggling to see how that demo would care about a theft of trade secrets case.
Whatever executive decision-making thread led to this being released needs to be recalibrated. I'd expect >13 year old maturity from a company even 1/10 the size of OpenAI.
OpenAI needs mass consumer buy in because a large part of their strategy is ad supported consumer use. Bad pr pushes those users away. A lot of people have teenage Reddit user mindset
I'm struggling to imagine someone like that who also cares about the minutiae of this case. Or heck, who even knows this case exists.
"OpenAI is so inept at system isolation that it accidentally hacked Huggingface" seems like a more important PR fire to put out. (Whether or not that breach was artificially manufactured)
Full self driving teslas on the moon any day now. AI data centers in space! We’re all going to be rich! What a whacked out timeline where this is the richest man.
Exactly. This reads like someone decided PR firms don't know what they're doing and decided to handle it themselves.
But... it turns out that professional people and consulting companies who have years of experience in a subject area might have something to offer. Who knew?!
Not to that extent though. For OpenAI the PR investment is disproportionate.
Regular companies have people that talk to the media or maybe run some ads. OpenAI is trying to have in-house media.
For a frontier lab it’s striking just how much energy (and dough) goes into this.
You’re right of course that PR exists elsewhere, but that obviously true assertion misses the finer details of how PR works at OpenAI. In doing so it functions as a thought-terminating cliché.
From: Che Chang <redacted>
Date: Feb 23, 2026, at 8:04 PM
Subject: Re: Former Apple Employees at OpenAI Retaining Non-public, Confidential, and Proprietary Information
To: <redacted>
Hi [Apple in-house legal counsel] and [Apple in-house legal counsel],
This guy from Weil claims to be representing you. For some reason, he is lying about speaking with me on the phone. I don't know who he is and we have never spoken.
I don't know what the point of his lie is, but could you please confirm if he is actually representing you, and if so, whether you have other counsel you can use. I don’t think interacting through him is appropriate given this behavior.
---
So, a lawyer emailed the wrong person. That's it. And they use that as a peg to hang their whole case on. I guess GPT 6 thought this was the real smoking gun!
Direct and unrehearsed doesn't seem like the right tack for a high-stakes legal dispute between two (allegedly) trillion-dollar companies. Especially since they've not been able to make a coherent point here.
Yes, we've built a society where being two-faced and false is rewarded. The least we can do is temper our criticism when someone "fails" to adhere to that.
I get what you're saying, but this seems like the wrong vehicle. I mean these are two corporations engaged in a huge and expensive legal battle—we're not talking about a personal he-said/she-said spat on X.
You got your crude analogy backwards - it is the calculated, tactical PR-speak that is more like shitting up the public square, than the naive, honest-if-self-serving speech.
Did no one ever explain to you why you should wear nice clothes to a job interview? Or a wedding?
Not simply that you should, but why? Conducting onesself with decorum and professionalism in certain contexts is not about deception. It's about respecting other people, both individually/directly, and the communal space.
Even if you consciously don't care about or respect anyone else, "everyone else is wrong for not being be ok with my t-shirt and flipflops" it's still simply a self-own to present this image of someone who lacks such a minimum level of perception.
A business communique like this isn't a performance art piece where it's fine if most people react badly to it because the whole point was to make people think about their reactions or reflect on some aspect of life. It's not the time or place to make a chocolate Jesus.
It's just some infants in pants they don't know how to fill, not "refreshing honesty".
Is that what we're calling PR-department-written say-nothing corporate lawyerspeak? "Decorum"?
I guess when people get banned, and only get cited a vague "violation of our terms of service", they should be grateful for how much professionalism and respect they're being shown? And actually being told what you are accused of is equivalent to wearing flip-flops to a wedding?
They need to get sued _hard_ and lose millions, as well as have cops put in jail before real change happens. I feel sorry for this kid but he's doing the right thing.
Taxpayers will just foot the bill. These cops need their personal funds taken, and before they blow it all on legal bills defending themselves from criminal charges.
Come to think of it, this is America dammit. Cops should be licensed, bonded and insured, and if their training or records aren't up to standards they don't get covered. We have higher standards for plumbers and pest control than we do police.
This. If truck drivers have to maintain safety records, if firefighters require certification and training, cops do too. To hell with this implied immunity bullshit, either they are correct in enforcing the law or they aren’t.
If it's one cops fault, charge the cop. If it's multiple cops fault, charge them AND the department. This makes it fiscally prudent to make sure cops in the department are well trained and educated on enforcing the laws and respecting rights.
I don't get the taxpayer line. It's a democracy. The taxpayers are the ones deciding what kind of policing we get.
If they don't want to foot the bills for police misconduct, maybe the taxpayers should elect someone who actually holds the police accountable and doesn't just blindly approve any policing budget and look the other way when the police wrongfully harms someone?
Right. The real justice comes from treating this incident like you would if there were no municipal coffers at play, and simply punish criminals for criminal action.
Real justice when it comes to swatting is multi-faceted.
Sure, the person making the bogus call to police should suffer some consequences vis a vis jail time.
However, swatting only works because police regularly take a shoot first and ask questions later stance. They go from an anonymous tip about an ongoing crime at an address to shooting up said address in short order without any effort to verify a crime is actually taking place, perhaps in no small part due to the fact that qualified immunity ensures they'll never see any consequences for any property damage and/or casualties which occur as a result.
I'd liken it to initiating a gang hit, except at least in the gang hit, the gang members responsible for carrying it out have a chance at facing some accountability.
Cops need to face the possibility of (at minimum) termination and even jail time for recklessly endangering civilians, and cop training needs to be fundamentally changed to eliminate this "us vs. them" mentality seen in so many of those in blue.
I am as horrified by the outcome here as the next person, but what exactly do you imagine police should do when responding to a call about an active shooter where they encounter a suspect armed with a handgun?
They could start by verifying the address of the house they're raiding. Way too often this doesn't happen and the police murder someone because they got the wrong house. They should identify the occupants of the house. The victim in this case had been swatted before and he had already warned the police that could happen. The swat team should have been skeptical going into this situation and taken the time to verify what was going on first. The call they got in this case claimed to be from the victim who said his father had shot his sister and he was hiding in another room. Even if the 911 call had been legit, the police would have shot the person who called for help and not the father, because they didn't bother to identify who they were shooting or who else was around when they sprayed bullets around the room from outside. They should have identified themselves as police. There's no reason not to. They should have at least attempted to use their big boy words before they started shooting up the place.
Identify themselves as LE, and of course physically represent as such, unambiguously, then demand the person put down the firearm. "Active Shooter" has no meaning just because the word is used. Also consider what everyone long forgot and few pay attention to; officers are paid, not volunteers. Risk is part of the job. You cannot shoot everything you think might be a threat. One can also question the quality of the system that handles emergency calls, or that responds to them. There are lots of guns in America. And lots of people who can say the words "active shooter". And often enough, phones and other objects are conveniently interpreted as guns, so I am not convinced that is the most important variable. Did he point it at the officers? If so, were the officers in proper uniform?
Now, if I were to answer your question, I would say it first depends on circumstances. I have encountered many people with guns. But if I was and active officer and encountered someone with a gun, I would do my best so that the encounter was arranged in a manner that gave me some advantage and at least a second or two to think before acting. Sometimes, mistakes are made though, and they always will be made. But the fact someone had a gun is no excuse.
The very smart and obvious thing to do would be to announce themselves loudly on arrival; all squad cars have very loud speakers on them and most police departments have at least one armored vehicle specially for high risk situations. Breaking in through a bathroom window and then running around the house with guns drawn is a recipe for disaster. If this kid had had good pistol skills (not that unusual unusual in the US) it might just as easily have ended with one or both of the cops getting shot.
Police in many jurisdictions will just walk onto your property in pursuit of a suspect and get mad at you if you open a window or door to see what the hell is going on outside your home.
This guy needs to be compensated for his pain and suffering, his economic losses due to his disability, and for the care he'll need for the rest of his life. Jailing the people whose negligence is responsible for his injuries doesn't help him.
There's quite a bit more negligence going on here than a typical murder—unless you've got some information that the cops were out to intentionally harm the victim.
The public would probably be served by the shooters not carrying a badge anymore, but it doesn't take jail to accomplish that.
The studies that found that your brain develops until you're 25 didn't study anyone older than 25.
More recent studies have found that your brain continues developing over your whole life. Different parts improve and decay at different rates at different times in your life.
There is no one age where you're ontologically perfect; it's all spectra all the time.
Just have an agent chug on a side-project for you, or set up a CI script to review every pull request or some similarly “helpful” task. That should eat a lot of tokens!
Man. If I had this kind of mandate I could really burn some tokens. Review each new PR and extract 100 topics to debate related to it. Spin up 1000 sub agents, each with a different personality profile system prompt, to debate each point until consensus has been reached. Synthesize the learnings into a limerick. Build a Spotify playlist that pairs with the tone of the debates. Post the limerick and link to playlist on the PR and tag me to notify me that I have a PR to review.
Oh man, when MCP was still new and shiny I made an MCP that let the AI choose appropriate theme music for what it was doing and it was an absolute blast, I need to make a more modern one.
Peer Gynt Suite's "In the Hall of the Mountain King" made a prominent appearance, but so did Aqua's "Barbie Girl"
Depends on how much you're asked to burn I guess. Until one point, it's actually helpful. Then there's a point where you can do stuff that's semi helpful but doesn't get in the way. But then I would imagine you reach a point where you have to come up with a token burn strategy and some kind of narrative for your manager that's in line with it. I bet at that last point, it gets taxing.
Probably like eating. Having to eat less to loose weight isn't great. Eating anything you want without worries is great. Having to eat more than you want to gain weight, not great.
you can't seriously think there is a line in the sand between homo sapiens and heidelbergensis or neanderthals which demarcates "soul" and "no soul". Especially since the latter interbred with humans.
Are there any species you know of that would be able to interbred with each other? Why would that have been possible?
Sounds like humans to me, perhaps ones who could live longer?
Look, I think we know how these discussions go, I'm not sure they get anywhere. Your pretty sure you know what's going on, I'm pretty sure I know what's going on...
Id hope there would be caution in manipulating the formation of a human when there isn't a great way to test these things without observing the results after they are born without any significant confidence what was manipulated effected the outcome or had a desirable result. There is no world where this real world testing is not reckless.
what if heidelbergensis and other ancestors didn't exist but were written in a books by people you never met and any amount of questioning was frowned upon? do you believe in the relics never seen and take the word upon a class of believers that safeguard the history? sounds like a religion to me.
reply