That would require replacing all the satellites with new ones capable of doing that, which doesn't seem feasible. Starlink also doesn't have great coverage of the polar regions.
“We're passing over terabits per second [of data] every day across 9,000 lasers,” SpaceX engineer Travis Brashears said today at SPIE Photonics West, an event in San Francisco focused on the latest advancements in optics and light. "We actually serve over lasers all of our users on Starlink at a given time in like a two-hour window.”
Again though, you can't do "Starlink from the satellites, so we don't rely on a specific ground station" unless all of the satellites support talking to Starlink, which they don't. That means they'd have to be replaced.
As I understand it, the Starlink network has a number of ground stations, and an active inter-satellite "mesh," thanks to laser links, which would allow it to route around the loss of one or more ground stations? (although obviously it requires at least one ground station to be live in order to access the non-Starlink-connected Internet)
The lasers began being integrated between 2020 and 2021, so it's likely SpaceX has made decent progress equipping their network with this capability, although I can't find the latest figures for the proportion of satellites that have lasers.
It sounds like there's something I'm missing if we can't do "Starlink from the satellites, so we don't rely on a specific ground station"
Do you understand what the problem trying to be solved is?
There are satellites in orbit today that have nothing to do with Starlink. Some of these have been up for a long time. We're talking weather satellites and research satellites. The ones in a polar orbit can only use one of two ground stations to communicate back with the earth, simply due to their location. One of those ground stations has lost it's fiber optic connection so it can't be used at full bandwidth right now.
None of that so far has anything to do with Starlink. We're talking about a system os satellites that already exists and predates Starlink sometimes by decades.
The person that started this thread proposed: "Maybe just use Starlink from the satellites, so we don't rely on a specific ground station." In other words, have these already existing satellites integrate with the (relatively new) Starlink system.
So they're saying that we somehow make those old satellites, which are already in orbit and have their own communication systems designed to interface with ground stations, somehow stop using the ground station and start using Starlink instead.
>> We're talking backup vs. primary. Of course the backup is not going to be as good.
Then it isn't really a backup. A lower-bandwidth failover capacity is properly described as an alternative or degraded pathway. To be a proper "backup" a thing has to actually do the primary job at least temporarily.
aye. Starlink could be, best case, an Out of Band (OOB) management interface.
good for getting into the other side of a connection or doing some management tasks like back-up telemetry -- but we're talking SNMP, SSH connections to routers, etc, not GigE levels of data.
Agreed. We really need some sort of regulation that prevents companies from bricking devices they sold. It's just so unethical, and wasteful, to put these devices out there and then just turn them into trash.
I do this technique too, and it's a great time for it. The OLED screen on the new iPad signals that Apple devices are moving to a better panel. If you've been waiting for the right time to move off an Intel Mac and onto a SoC Mac, it's now. Pick up a refurbished M2 MacBook. They're in the sweet spot for support, power, and cost.
The next one will probably have an OLED screen; so if you wait til then, your refurb M1/2/3 will be on Apple's short list of devices they don't want to support. (And you might have panel FOMO.) Or you'll have to pay the premium price for the latest model.
If this really was a mistake the easiest way to deal with it would be to release people from their non disparagement agreements that were only signed by leaving employees under the duress of losing their vested equity.
It's really easy to make people whole for this, so whether that happens or not is the difference between the apologies being real or just them just backpedaling because employees got upset.
Edit: Looks like they're doing the right thing here:
> Altman’s initial statement was criticized for doing too little to make things right for former employees, but in an emailed statement, OpenAI told me that “we are identifying and reaching out to former employees who signed a standard exit agreement to make it clear that OpenAI has not and will not cancel their vested equity and releases them from nondisparagement obligations” — which goes much further toward fixing their mistake.
This reads like more than standard restrictions. I hate those like everyone, they are just intended to chill complaints in my opinion with enough question to scare average people without legal expertise (like me, like most devs), just like non-competes used to seemingly primarily be used to discourage looking at other jobs, separate from whether it was enforceable - note the recent FTC decision to end non-competes.
About 5 months ago I had a chance to join a company, their company had what looked like an extreme non-compete to me, you couldn't work for any company for the next two years after leaving if they had been a customer of that company.
I pointed out to them that I wouldn't have been able to join their company if my previous job had that non-compete clause, it seemed excessive. Eventually I was in meetings with a lawyer at the company who told me it's probably not enforceable, don't worry about it, and the FTC is about to end non-competes. I said great, strike it from the contract and I'll sign it right now. He said I can't do that, no one off contracts. So then I said I'm not working there.
I have worked for multiple startups (Malwarebytes, Vicarious, Rad AI, Explosion AI, Aptible, Kenna Security). Not once have I seen an exit agreement that stated they would steal back my vested equity if I didn't sign. This is definitely not "standard restrictions".
Yes, I really do this. Have done since I started working.
At one of my first jobs as a student employee they offered me a salary X. In the contract there was some lower number Y. When I pointed this out, they said "X includes the bonus. It's not in the contract but we've never not paid it". OK, if this is really guaranteed, you can make that the salary and put it in writing. They did, my salary was X and that year was the first time they didn't pay the optional bonus. Didn't affect me, because I had my salary X.
IANAL and I don't know how binding this is. I'd think it's crucial for it to be in both copies of the contract, otherwise you could have just crossed it out after the fact, which would of course not be legally binding at all and probably fraud (?)
In practice, it doesn't really come up, because the legal department will produce a modified contract or start negotiating the point. The key is that the ball is now in their court. You've done your part, are ready and rearin' to go, and they are the ones holding things up and being difficult, for something that according to them isn't important.
UPDATE:
I think it's important to note that I am also perfectly fine with a verbal agreement.
A working relationship depends on mutual trust, so a contract is there for putting in a drawer and never looking at it again...and conversely if you are looking at it again after signing, both the trust and the working relationship are most likely over.
But it has to be consistent: if you insist on a binding written agreement, then I will make sure what is written is acceptable to me. You don't get to pick and choose.
For one job I also crossed some stuff out, founder was cool with it because he mostly got it from a template. Having actual paper is great for that. At a later job, they insisted on Docusign, which was basically I get an immutable image in a browser to 'electronically sign' with no modifications. It had a section that amounted to a non-compete agreement that I didn't like, but their lawyers didn't really answer me on whether such a thing could be enforced or not given that the company was headquartered in California even though I'd be working out of Washington. I took that as a sign that they probably wouldn't go Amazon on me, at least.
First you’re offering up a lot of trust to people you might have just started working with.
Or, they could be very trustworthy and just remember things differently. And of course people come and go in companies all the time they just might not be there.
At least if you do a verbal agreement follow it up with an email confirming the details.
Exactly. And just like I have to be fine with not getting the job if my conditions are not acceptable to them, they have to be fine with not getting me if their conditions are not acceptable to me.
Considering the considerable effort that has gone into this by the time you are negotiating a contract, letting it fail over something that "is not important" and "is never enforced" would be very stupid of them.
So if they are unwilling to budge, that either means they were lying all along and the thing that's "never enforced" and is "not important" actually is very important to them and definitely will be enforced, or that they are a company that will enforce arbitrary and pointless rules on employees as long as they think they can.
Neither of which is a great advertisement for the company as an employer.
> So if they are unwilling to budge, that either means they were lying all along and the thing that's "never enforced" and is "not important" actually is very important to them and definitely will be enforced, or that they are a company that will enforce arbitrary and pointless rules on employees as long as they think they can.
Most of the time is basically just FUD, to coerce people into following the rule-that-is-never-enforced
It's always "not enforced" or "just a template" right up until they decide they need to pressure you into something and then they will have no problem referencing those items.
Do not sign a contract unless you are willing to entirely submit to everything in it that is legally binding.
Also be careful with extremely vague contracts. My employment contract was basically "You will do whatever we need you to do" and surprise surprise, unpaid overtime is expected.
I've seen legal departments redlining drafts of a contract repeatedly until an agreement had been reached. The final contract still contained the red lines.
IANAL but I've seen strikes throughout contracts, and then an initial+date from both parties. Weird how in 2024 an initial that's so easily forgeable can be legally binding
I guess like all laws it depends on jurisdiction, and more importantly, if you can convince the judge/magistrate that the contract did or did not happen
Yeah, but the law still sets the default baseline for he judge / jury / whatever. In the Nordic countries (at least Sweden) -- as techno-modern and bureaucratic as they may be -- that still includes verbal agreements.
(The handshake is probably not a legal requirement, though I suppose it could be taken into consideration as evidence -- "You even shook hands on it, so you must have realised that what you had just discussed were atually the terms you were agreeing to.")
By that downplaying the significance of a contract's terms and persuading to sign it without fully understanding the implications. If this were only the beginning of my career, meaning my first serious company, I would sign the contract.
I’ve heard of some pretty aggressive non-competes in finance, but AFAIU (never worked in Connecticut myself), it’s both the carrot and the stick: you get both paid and a stiff contract if you leave with proprietary alpha between the ears.
In tech I’ve never even heard a rumor of something like this.
It’s got a term - “garden leave” and yeah it was prevalent in finance. I say “was” because I think some states are changing laws wrt/ non-competes and this calling this practice into question.
First of all, taking any code with you is theft, and you go to jail, like this poor Goldman Sachs programmer [1]. This will happen even if the code has no alpha.
However, noone can prevent you from taking knowledge (i.e. your memories), so reimplementing alpha elsewhere is fine. Of course, the best alpha is that which cannot simply be replicated, e.g. it depends on proprietary datasets, proprietary hardware (e.g. fast links between exchanges), access to cheap capital, etc.
What hedge funds used to do, is give you lengthy non-competes. 6months for junior staff, 1-2y for traders, 3y+ in case of Renaissance Technologies.
In the US, that's now illegal and un-enforceable. So what hedge funds do now, is lengthy garden(ing) leaves. This means you still work for the company, you still earn a salary, and in some (many? all?) cases also the bonus. But you don't go to the office, you can't access any code, you don't see any trades. The company "moves on" (developes/refines its alpha, including your alpha - alpha you created) and you don't.
These lengthy garden leaves replaced non-competes, so they're now 1y+. AFAIK they are enforceable, just as non-competes while being employed always have been.
I think this still leaves garden leave on the table. The thing that can no longer happen is an employer ending it's relationship with an employee and preventing them from continuing their career after the fact. Garden leave was in fact one of the least bad outcomes of a non-compete as I understand it.
Comp clawbacks are quite common in finance, at least contractually. It's rare for it to go ahead, but it happens. It isn't some especially weird thing.
Comp clawbacks in exit agreements, that weren't part of the employment agreement?
I've seen equity clawbacks in employment agreements. Specifically, some of the contracts I've signed have said that if I'm fired for cause (and were a bit more specific, like financial fraud or something) then I'd lose my vested equity. That isn't uncommon, but its not typically used to silence people and is part of the agreement they review and approve of before becoming an employee. It's not a surprise that they learn about as they try to leave.
It must have been part of the original employment document package, that the equity was cancellable. In the details of the equity grant, or similar, somewhere.
According to the Vox article, it's much more complicated legally. It's not part of each employee's contract that allows this, it's part of the articles of incorporation of the for-profit part of OpenAI.
Do you mean it is generic to do that in contracts? (Been a while since I was offered equity.)
Or do you mean that even OpenAI would not try it without having set it up in the original contract? Because I hate to be the guy with the square brackets ;-)
If it wasn't in the original contracts for the equity, they wouldn't be able to claw back.
Fairly obviously, the mechanism can't be in the exit agreement because you didn't sign that yet.
Normally a company has to give you new "consideration" (which is the legal term for something of value) for you to want to sign an exit agreement - otherwise you can just not bother to sign. Usually this is extra compensation. In this case they are saying that they won't exercise some clause in an existing agreement that allows them to claw back.
Per the Vox article, it's not directly in the contract you sign for the equity, it's basically part of the definition of the equity itself (the articles of incorporation of the for-profit company) that OpenAI remains in full control of the equity in this way.
What is the structure of those compensations, and the mechanism for the clawbacks? Equity is taxed when it becomes the full, unrestricted property of the employee, so depending on the structure these threatened clawbacks could have either (1) been very illegal [essentially theft], or (2) could have had drastic and very bad tax consequences for all employees, current and former.
I'm not surprised that they're rapidly backpedaling.
> taxed when it becomes the full, unrestricted property of the employee
I guess these agreements mean that the property isn't full unrestricted property of the employee... and therefore income tax isn't payable when they vest.
The tax isn't avoided - it would just be paid when you sell the shares instead - which for most people would be a worse deal because you'll probably sell them at a higher price than the vest price.
It's a worse deal in retrospect for a successfull company. But there and then it's not very attractive to pay an up-front tax on something that you can sell at an unknown price in the relatively far future.
Not sure how they deal with the tax. Ping John Stumpf (former Wells CEO) and ask, he probably has time on his hands and scar tissue and can explain it.
Common? Absolutely not. It might be common for a tiny fraction of investment bank staff who are considered (1) material risk takers, (2) revenue generators, or (3) senior management.
Can you find any specific examples? I've only seen that apply to severance agreements where you're being paid some additional sum for that non-disparagement clause.
Never seen anything that says money or equity you've already earned could be clawed back.
Right, but would that have been achieved with a clause open-ended enough to allow this additional paperwork on exit?
Or would that have been an "if you break the law" thing?
Seems unlikely that OpenAI are legally in the clear here with nice clear precedent. Why? Because they are backflipping to deny it's something they'd ever do.
I think they are backpedaling rapidly to avoid major discontent among their workers. By the definition of their stock as laid out in their articles of incorporation, they have the right to reduce any former employee's stock to 0, or to prevent them from ever selling it, which is basically the same thing. This makes their stock offers to employees much less valuable than the appear at face value, so their current and future employees may very well start demanding actual dollars instead.
> The ability to do it was almost certainly already in place
Why? OpenAI is a shitshow. Their legal structure is a mess. Yanking vested equity on the basis of a post-purchase agreement signed under duress sounds closer to securities fraud than anything thought out.
I'm not saying it was thought out, I'm saying it was in place. My understanding is that the shareholders agreement had something which enabled the canceling of the shares (not sure if it was all shares, shares granted to employees, or what). I have not seen the document, so you may be right, but that's my understanding.
> the shareholders agreement had something which enabled the canceling of the shares
OpenAI doesn't have shares per se, since they're not a corporation but some newfangled chimeric entity. Given the man who signed the documents allegedly didn't read them, I'm not sure why one would believe everything else is buttoned up.
If its not negotiated on exit why are they requesting additional documents to be signed when leaving? Clearly nothing like this was agreed at the start of employment.
Would it though? Presumably a finance company's claw back clause is there to protect it from you taking its trade secrets with you to its competitors, not from you tweeting "looks trashy lol" in response to a product launch of theirs, or you mentioning to a friend that your old boss was kind of a dick.
I’ve seen that for a well-known large tech company, and I wasn’t even employed in the US, making those seem stranger. Friends and former colleagues pushed back against that (very publicly and for obvious reasons in one case) and didn’t get to keep their vested options: they had to exercise what they had before leaving.
There was one thing that I cared about (anti-competitive behavior, things could technically be illegal, but what counts is policy so it really depends on what the local authority wants to enforce), so I asked a lawyer, and they said: No way this agreement prevents you from answering that kind of questioning.
Often RSUs in non public companies come with a double trigger: you need both the vest to happen and a liquidity to happen for the actual delivery of those, so no tax implications until a liquidity event (afaik. But don’t take tax advice from randos on the internet).
The closest thing I've heard of is having to sign anti-disparagement clauses as part of severance when laid off; still pretty shitty, but taking back already vested equity would be on another level.
My understanding is that its an explicit condition of the equity grant, not something technically first revealed at exit (which would probably be illegal), but probably under the expectation that no one is carefully studying the terms of the agreement that would be required at exit when they are accepting compensation terms that i nclude equity.
> I have worked for multiple startups (Malwarebytes...
Note the "have worked" and the rather long list of places they've worked. If that list is in chronological order (sure didn't look alphabetical), Malwarebytes doesn't have to be a startup now for it to have been one when GP worked there.
"That's alright, as you said it's not enforceable anyway just remove it from everyone's
contract. It'll just be the new version of the contract for everyone."
Doubt it would have made any difference though, as the lawyer was super likely bullshitting.
> I was in meetings with a lawyer at the company who told me it's probably not enforceable, don't worry about it
Life rule: if the party you're negotiating a contact with says anything like "don't worry about that, it's not enforceable" or "it's just boilerplate, we never enforce that" but refuses to strike it from the contract then run, don't walk, away from table. Whoever you're dealing with is not operating in good faith.
You did well: there is never a rule against one-off contract. I can assure you the CEO has a one-off contract, and that lawyer has a one-off contract, at the very least :D
Those are great points. I didn't think about it at the time. Since they were pushing me hard to sign a contract that for once literally blocked most of the companies in the cs speciality area I mostly have worked in, it gave me enough courage to say no. Barely ;-)
Yeah, totally legit. Don't worry about it, it's not enforceable anyways. What, remove it from the contract? God no! Oh, I mean sorry, no one off contracts.
Well, you can be surprised. It's surprisingly common, in my experience, to believe people who pretend they are on your side. One interesting and typical case that is documented through countless online videos is police interrogations, where the interrogator is usually an expert in making it seem he (or she) is on your side, despite how obvious it should be that they're not. "Can I get you a meal?", friendly tone, various manipulations and before you know it you've said things that can and will be used against you whether you are guilty or not.
> We can also mention the case of psychiatrists running the "Presence francaise" groups who, appointed to examine the prisoner, started off boasting they were great friends with the defense lawyer and claiming both of them (the lawyer and the psychiatrist) would get the prisoner out. All the prisoners
examined by this method were guillotined. These psychiatrists boasted in front
of us of this neat method of overcoming "resistance."
Attorneys are like any other profession. The average attorney is just like the average person, except he passed a difficult test.
Exceptions require sign off and thinking. The optimal answer is go with the flow. In an employment situation, these sorts of terms require regulatory intervention or litigation to make them go away, so it’s a good bet that most employees will take no action.
Probably not enforceable != enforceable. Are you worth suing or does everyone sign? Are your state laws and jurisprudence going to back you up?
If you are ever going to sign an employee agreement that binds you, consult with an employment attorney first. I did this with a past noncompete and it was the best few hundred I ever spent: my attorney talked with me for an hour about the particulars of my noncompete, pointed out areas to negotiate, and sent back redlines to make the contract more equitable.
The single best professional decision I ever made was to get a business degree. The degree itself wasn’t worth a damn, but the network was invaluable. I have very close friends who are the exact kind of attorney who you would expect to have an undergraduate business degree. They’re greedy, combative people who absolutely relish these sorts of opportunities. And as a bonus, they are MY greedy, combative people who relish these sorts of opportunities.
They’re great partners when confronted with this kind of contract. And fundamentally, if my adversary/future employer retains counsel, I should too. Why be at a disadvantage when it’s so easy to pay money and be at even?
There are some areas my ethics don’t mesh with, but at the end of the day this is my work and I do it for pay. And when I look at results, lawyers are the best investment I have ever made.
At most it would just make that part of the contract void. Almost all contracts with stuff like this would have a “severability” clause which states like if one part of the contract is invalid, the rest is still valid.
But even without that, judges have huge amounts of leeway to “create” an ex post facto contract and say “heres the version if that contract you would have agreed to, this is now the contract you signed”. A sort of “fixed” version of the contract.
> At most it would just make that part of the contract void. Almost all contracts with stuff like this would have a “severability” clause which states like if one part of the contract is invalid, the rest is still valid.
Severability clauses themselves are not necessarily valid; whether provisions can be severed and how without voiding the contract is itself a legal question that depends on the specific terms and circumstances.
There is usually a severability clause that basically says if a clause is illegal it voids that clause not the whole contract… this is pretty standard practice.
The article makes it clear that it wasn't a mistake at all. It's a lie. They were playing hardball, and when it became public they switched to PR crisis management to try and save their "image", or what's left of it.
They're not the good guys. I'd say they're more of a caricature of bad guys, since they get caught every time. Something between a classic Bond villain and Wile E. Coyote.
"...But there's a problem with those apologies from company leadership. Company documents obtained by Vox with signatures from Altman and Kwon complicate their claim that the clawback provisions were something they hadn’t known about..."
"...and that our PR firm wasn't good enough to squash the story."
They will follow the standard corporate 'disaster recovery' - say something to make it look like they're addressing it, then do nothing and just wait for it to fall out of the news cycle.
Honestly I'm willing to give the benefit of the doubt on that, depending on their actions, because I'm sure they sign so many documents they just rely on their legal teams to ensure they're good.
There's absolutely no way that the officers of the company would be unaware of this.
First of all, it beggars belief that this whole thing could be the work of HR people or lawyers or something, operating under their own initiative. The only way I could believe that is if they deliberately set up a firewall to let people be bad cops while giving the C-suite plausible deniability. Which is no excuse.
But...you don't think they'd have heard about it from at least one departing employee, attempting to appeal the onerous terms of their separation to the highest authority in the company?
Hold up... Do you really think that a C-suite including career venture-capitalists who happen to be leading+owning stock in a private startup which has hit an estimated billion+ valuation are too naive/distracted to be involved in how that stock is used to retain employees?
In other words, I'm pretty sure the Ed Dillingers are already in charge, not Walter Gibbs garage-tinkerers. [0]
"Trust me bro, if it weren't up to me you wouldn't even have to sign that contract. I mean it is up to me, but, like, I won't enforce the thing I made you sign. What? No I won't terminate the contract why don't you trust me bro? I thought we were a family?"
Yeah, agree, but they don't have to cancel the disparagement clause. They could just eat the PR hit. Allowing former employees to talk freely seems risky to me (if I were them). I think we can give them back 5 points for this move but still leave them at -995 overall.
For whatever it's worth (not much), Sam Altman did say they would do that
> if any former employee who signed one of those old agreements is worried about it, they can contact me and we'll fix that too. very sorry about this.
What Sam is saying is very different than what I'm saying. I'm saying he should be proactive and just do it, he's saying that if people explicitly reach out to him then he'll do it specifically for them.
Importantly if he just said publicly that that he wouldn't enforce the non-disparagement agreements that could be legally binding[1]. But if he just says he'll release people who he's, legally speaking, free to just not do that.
[1] The keywords are promissory estoppel. I'm not a lawyer but this looks at least like a borderline case worth worrying about.
Sure and anyone who has worked in a toxic workplace knows exactly what it means to require a direct path to leadership to resolve an issue instead of just resolving it.
True, they can't renegotiate agreements that don't yet exist.
However the fact that the corporate leadership could even make those threats to not-yet-departed employees indicates that something is already broken or missing in the legal relationship with current ones.
A simple example might for the company to clearly state in their handbook--for all current employees--that vested shares cannot be clawed back.
Given what OpenAI's been in the press for the past few weeks, I can't help but feel this is a trap; even if it isn't, Sam is certainly making it look like it is...
Some people are just so deep into fanboiism they refuse to see the writing on the wall even when it's in ten-foot-high letters of fire on a pitch-black background. Of fucking course it's just a performance for the public.
That really is not enough. Now that they have been publicly embarrassed and the clause is common knowledge they really have to undo the mistake. If they didn't, they would look like a horrible employer and employees would start valuing their stock at $0, dropping their effective compensations by a ton and then people will leave. Given the situation, undoing the agreement is an act of basic self-preservation at this point.
The documents show this really was not a mistake and "I didn't know what the legal documents I signed meant, which specifically had a weird clause that standard agreements don't" isn't much of a defence either. The whole thing is just one more point in favor of how duplicitous the whole org is, there are many more.
> ”we are identifying and reaching out to former employees who signed a standard exit agreement to make it clear that OpenAI has not and will not cancel their vested equity and releases them from nondisparagement obligations”
Well, they say they are. But the nondisparagement agreement repeatedly forbids revealing the agreement itself, so if it wasn't cancelled those subject to it would be forbidden to point out that the public claim they were going to release people from it was a lie (or done only for people from whom OpenAI was not particularly concerned about potential disparagement.)
“disparagement" is whatever is defined in the agreement, which reportedly (from one of the people who declined to sign it) includes discussing the existence of the agreement.
This is not normal for being a corporate employee. This was certainly going to come out eventually and cause big problems, but to the extent Sam thinks AGI is around the corner he might not be playing the long game.
"We're hereby making a legally binding commitment that those clauses are void, whether anyone reaches out to us or we manage to reach out to them or not."
Unless and until that's what they say, looks like they’re not doing that.
> We're removing nondisparagement clauses from our standard departure paperwork, and we're releasing former employees from existing nondisparagement obligations unless the nondisparagement provision was mutual. We'll communicate this message to former employees.
So the former successfully blackmailed employees, stay blackmailed.
Even if that's true (and I'm not saying it is, or it isn't, I don't think anyone on the outside knows enough to say for sure), is it because they genuinely agree they did something egregiously wrong and they will really change their behavior in the future? Or is it just because they got caught this time so they have to fix this particular mistake, but they'll keep on using similar tactics whenever they think they can get away with it?
The impact of such uncertainty on our confidence in their stewardship of AI is left as an exercise for the reader.
> Edit: Looks like they're doing the right thing here:
> Altman’s initial statement was criticized for doing too little to make things right for former employees, but in an emailed statement, OpenAI told me that “we are identifying and reaching out to former employees who signed a standard exit agreement to make it clear that OpenAI has not and will not cancel their vested equity and releases them from nondisparagement obligations” — which goes much further toward fixing their mistake.
1. Get cash infusion from microsoft
2. Do microsoft playbook of 'oh I didn't mean to be shady we will correct' when caught.
3. In the meantime there are uncaught cases as well as the general hand waving away of repeated bad behavior.
4. What sama did would get him banned from some -fetish- circles, if that says something about how his version of 'EA' deals with consent concerns.
Plenty of legitimate things to criticize EA for, no need to smear them by association with someone who's never claimed to be an EA and hasn't obviously behaved like one either.
It shouldn't take a Vox article to ensure employees basic security over their compensation. The fact that this provision existed at all is exceptionally anti-employee.
You surely don't actually believe Altman when he says they're doing this? Like Elon Musk, Altman is a known liar and should not be trusted. It's truly unbelievable to me that people take statements like this at face value after having been lied to again and again and again. I think I'm starting to understand how crypto scams work.
Alphabet has to publish or they would lose a huge amount of their research staff. People in this industry want to publish and will work at the companies that allow it before they work at the companies that restrict it.
They aren't doing it out of the goodness of their heart, they're doing it because it benefits them to. If they want to have those benefits they should be held to the same standards as everyone else.
The market value (stock price) assumes a lot more growth, so the idea that volumes are dropping is pretty bad if your wealth is based on that stock price.
Markets also aren't zero sum- if the overall volume of EVs keeps rising then competitors entering the market don't necessarily mean your volume drops, as they could be taking a slice of the increase in overall sales.
Are you really claiming that perfluorooctanoic acid (PFOA) is safe? I'm not going to quote the whole "toxicology" section of wikipedia on that chemical but strongly recommend you read it:
PFOA was explicitly banned in the US in 2014, after being phased out the year before, due to it's toxicity. Teflon changed their manufacturing process so it's made with polytetrafluoroethylene instead. So yes, pans made with teflon before 2014 absolutely due contain a toxic chemical.
> Teflon changed their manufacturing process so it's made with polytetrafluoroethylene instead. So yes, pans made with teflon before 2014 absolutely due contain a toxic chemical.
Incorrect. Just because something is used as a chemical precursor in a manufacturing process, does not mean that it is present in the final manufactured product.
Teflon is an solid, non-reactive material. PFOA is a liquid acid. They are very, very different things. Even if manufactured Teflon contained any residual PFOA (it doesn't), it would be washed away the first time you rinsed the pan.
> Most manufacturers assumed that PFOA burns off during the process of manufacture, but traces of PFOA were found in some Teflon-coated cookware.
> A 1999 study found that 98% of people in the United States had PFOA in their blood. This was due to environmental exposure to the chemical. As a result, the US EPA put up a program to eliminate the use of PFOA by 2015.
They just assumed it burned off, but never tested it. It wasn't until 2021 that they had to actually test and report for extra chemicals left in, well after it was banned.
What happens when you actually test it?
> In this study, we identified and measured perfluoroalkyl carboxylates (PFCAs), particularly PFOA, and fluorotelomer alcohols (FTOHs; 6:2 FTOH and 8:2 FTOH), released from nonstick cookware into the gas phase under normal cooking temperatures (179 to 233 degrees C surface temperature). PFOA was released into the gas phase at 7-337 ng (11-503 pg/cm2) per pan from four brands of nonstick frying pans.
Hilariously, the webmd link you cited repeats what I said above:
> The use of PFOA in the manufacturing of Teflon-coated cookware has been completely stopped. But, even when PFOA is used, it poses little or no harm to your health. Teflon on its own is safe and can’t harm you when you ingest it.
...not that I think webmd is worth citing.
But OK, one paper you dug up found an incredibly tiny amount (500 picograms per square centimeter, at the highest), which they said declined over time for half the pans tested (I'm guessing the ones with the highest initial readings). Meanwhile, other groups didn't see the same thing:
> Commercial grade cookware was obtained, then extracted with water and ethanol/water mixtures at 100 and 125 degrees C, and the resulting extracts were analyzed by liquid chromatography tandem mass spectrometry (LC/MS/MS). Detection and quantification limits as low as 100 pg cm(-2) were demonstrated. None of the fluoropolymer treated cookware samples analyzed showed detectable levels of PFOA when extracted under simulated cooking conditions.
Anyway, I'm not going to get excited over a single paper, showing incredibly minute amounts of something which, as you say, is now ubiquitous in the environment (hello, cross-contamination risk!). The paper may or may not have been methodologically valid. You do you.
reply