In part, yes, but the whole picture is a bit more complex. Intel and AMD both work with vendors to work out reference designs and power regulator configurations that are compatible with their CPUs. Modern power regulators are relatively complex, especially when you get into the high-end desktop and server space. There are a lot of things that can be done wrong and cause issues, such as voltage or current overshoot when the CPU moves between different power states (e.g. low to high load).
OpenAI does not have an advisory board. She is on the board of directors and has fiduciary responsibilities. It does not matter that the parent organization is a non-profit. The duty remains.
Yes, those responsibilities are to advance the charter of the organization, which notably is totally compatible with publishing articles that contain mild criticism of the organization's behavior.
The board’s fiduciary duties are to the charter of the nonprofit. They have no duty to protect the for-profit-OpenAI’s profits, reputation, or anything else.
This has been argued here already and dissected thoroughly.
It's basically restating 'her role is to cheerlead' but with more sophisticated wording.
You haven't made an argument, you've just used longer words to restate your (incorrect) opinion.
Also I never said they had an 'advisory' board. Board members ar either executive or non-executive members. If they're nonexec then their role is to advise the executive members. They're generally fairly independent and many nonexecutive board members serve on the boards of multiple companies, sometimes including ones that are nominally in competition with each other.
No, I never said that her role is to "cheerlead". Refraining from public disparaging remarks is not a big ask for a board member. Especially if those thoughts were not first brought before the board, which is where such disputes should be resolved.
> Also I never said they had an 'advisory' board
From your comment: "...it's absolutely the role of advisory board members to continue their work".
Not sure how someone is meant to parse that except for you to imply that she was a member of an advisory board.
Someone is meant to use context and understand that when the context clearly refers to someone who is well known to be a member of the board of directors, 'advisory board member' means that they are a board member in an advisory role, aka a nonexec.
That assumes that someone is familiar with the structure and makeup of boards of directors. It's a bit rude to participate in a conversation if you're not familiar with the subject matter and the meaning of words-in-context as you just waste peoples time making them elaborate.
> 'advisory board member' means that they are a board member in an advisory role, aka a nonexec
No, those are not synonyms. Advisory boards are distinct from the board of directors. Advisory board members have roles similar to what you alluded to in your original comment and do not have fiduciary duties, hence my confusion.
Non-exec board members are not involved with day-to-day business operations but their fiduciary duties are no different than exec members.
Language has meaning. Don't insult others for your own clumsy usage.
Ehh I don't think SVB is an apt comparison. When the FDIC takes control of a failing bank, the bank shutters. Only critical staff is kept on board to aid with asset liquidation/transference and repay creditors/depositors. Once that is completed, the bank is dissolved.
While it is true that the govt looks to keep such engagements short, SVB absolutely did not shutter. It was taken over in a weekend and its branches were open for business on Monday morning. It was later sold, and depositors kept all their money in the process.
It seems that your primary concern is that the government (or some bad actor) will be able to install a backdoor into PQC algorithms. Is that right? Why would PQC be more exposed to this type of subversion than existing public-key cryptography?
To your point about PQC being used exclusively, post-quantum encryption methods are designed to be resistant to both quantum and classical attacks. That is one of the key stated goals of the NIST post-quantum cryptography program.
it's less about a backdoor and more about just being a lot less robust in general. classical crypto is based on ~100 years of math on finite fields and ~50 years of heavy scrutiny by cryptographers. the post quantum algorithms are much newer and built on much less well studied math. (and empirically, a large number of them have been found to be completely broken). we're at least 20 years from PQC that can be widely trusted. there really just isn't an alternative to having a generation of grad students studying an algorithm that's as old as they are
For signatures, hash based signatures are quantum computer resistant and are also more secure than any other signature scheme. No reliance on a math problem if you don't count the cryptographic permutation to be one, but then everything relies on it regardless of what scheme is used.
The McEliece cryptosystem[1] is one of finalists in the PQC competition and it's also quite old - developed in 1978. It didn't face as much scrutiny as RSA or ECC due to its large key sizes which resulted in nonexistent adoption.
My understanding is that all the other PQC candidates including Kyber are much newer and far less studied.
High-end CPUs are actually not powered directly by DC. Basically all server (and growing portion of consumer) CPUs are powered by multiphase buck regulators[1] which split the power from from the DC PSU rails into a parallel set of modulated buck regulator power stages. The outputs of the parallel regulators are recombined to generate DC (as the combined waveform of the phased AC parts).
The reason for this multiphase design is because it offers better power efficiency and better transient response to the CPU as the CPU moves between it's different power states (high vs low load).
How would that work? If I just comment on some 8 year old question that the answer does not apply to $LATEST version of $FRAMEWORK, how would anyone with the know-how discover my question?
This truly is a new question because what is really being asked is what was the breaking change that occurred in $FRAMEWORK between the original question was answered and this latest version.
Before the node-size race we also had the clock-speed race. Eventually it was common for processors reach 2-4ghz, and after that the clock speed gains stopped being practical because as you increase clock speed you also increase energy requirements and heat.
I think the implication is that clock-speed could start increasing again. It would probably require a completely new manufacturing process, but if we assume this superconductor is legit, perhaps an older process could manufacture it.
If so, maybe we could have (just spitballing here, I have no idea) 28nm super conducting CPUs that run at a 1thz instead of 4ghz. That would be quite an improvement over today's CPUs, even with fewer transistors, I think.
There are other losses and limitation in increasing clock-speeds aside from just resistive losses, but I think they are a significant part of the current bottleneck. Other losses involve transistor switching losses, and inductive losses but I don't really know the details, and I think those details change with superconductors.
We stopped chasing clock speeds because of the physical timing limitations of gate and signal propagation. Not because of heat. Suppose you are using a 5GHz clock. Every cycle is 0.2ns. Light can only travel 6cm in that time. Electricity propagates a little slower through a conductor (and even slower through silicon). So if you are using some insanely fast clock, you are just wasting cycles waiting for signals to move across the chip.
Current processors are no longer synchronous, each part now works asynchronously and there can be several instructions waiting to be completed at once, average Intructions Per Clock are already over 1, so there is no problem waiting a little more for signal to propagate.
I agree with this sentiment. I have also been somewhat unsettled by the detractors here. I am not sure if this is purely an American sentiment, but it seems like more of the general population has started to view advancement in binary terms. In both science and public policy, it seems that an increasing proportion of the population view anything less than 100% success/improvement as an abject failure.
“ChatGPT couldn’t solve world hunger, nor could it come up with a peaceful solution to Israel and Palestine when I asked it! See, I told you it’s useless and not really intelligent.” — hyperbole but not even that far from some comments I’ve seen.
I don't know why you're getting downvoted, because this is precisely what I'm seeing from both the IT crowd and the general public.
"Computers understand English now, but they occasionally make mistakes, so who cares?"
Something I've noticed is that GPT 4 seems to make mistakes less often than humans. That is, if you asked a random person to stand up in a lecture hall and answer questions thrown at them like the type people have been trying on ChatGPT, they would fail at least 50% of the time, like more.
For example, from what I can tell, GPT 4 has nearly perfect spelling and grammar. Better than mine, certainly, and up there with professional copy editors.
> I have also been somewhat unsettled by the detractors here.
I wouldn't worry about it. The news papers always view it in a sorta pessimistic way. I used to joke about this with spacex's booster landings. When they couldn't stick a landing even though the mission itself was successful, the news sites always made headlines that sounded like "welp, spacex screwed up again" lol :D
The news sites will always pick the headline that raises the most eyebrows.
As for my liking of spacex but my great distaste for the jackass known as Elon, well, that's nothing new. I was saying that 10 years ago lol. You can still cheer on the accomplishments all of the engineers are spacex are making. That's fine.
The problem with it is this is actually burning many millions of US taxpayer dollars. You think Musk is paying for it? Check SpaceX grants.
As for the failure, the Raptor 2 failures have caused a RUD on multiple Starship launches so far.
They should be back to static firing the Super Heavy or something, but this is not sexy enough for the billionaire.
Until the engine and launchpad works reliably, there is literally no point in launching the boosters when you have high probability of mission failure right at the launch. (Even if it clears the tower.)
Musk/SpaceX redefining failure as success is terribly annoying too. If done repeatedly enough it has the potential to tank the whole space program. They obviously wanted to test separation the most (since it was issued despite the control failure) and that failed too.
The director of nasa calls this a great success. So does a former astronaut.
> Thursday's launch was hailed as "a real accomplishment" and "so successful" by NASA Administrator Bill Nelson and retired International Space Station Commander Chris Hadfield, respectively. SpaceX agreed.
I read it more as jeering at private companies who promise the world and deliver a fraction of the product with price-gouging tactics.
Somewhere in there is an innovation that might do a public good, but most people are realising that corporations don't do public good. They do what they want.
If it was a government entity that tried and failed to reach for the stars, I think there'd be jeering too ("there go my tax dollars") but a large portion of the scientific community would be happy with the result.
>I am not sure if this is purely an American sentiment
My observation is that it is an online sentiment. But due to the way Technology news cycle works most of these online sentiment do have their source from America. So I think it is an American Sentiment being exported. But most places outside America, especially those without the usage of English being main language tends to be less affected.
I'm neither a fan nor a detractor of Elon Musk. The man has his faults. I am a huge fan of space exploration and the settlement of Mars, and therefore I really, really want SpaceX to succeed. For that reason I've been very critical of this test.
Professionals in the industry whose life work is the study of rocket plumes (e.g. @ DrPhiltill on Twitter) warned about what would happen if they proceeded with this plan. They were right. The FAA license application, as it turns out, was wildly off base and the projected environmental impact was off by as much as an order of magnitude, being based on an earlier design with much, much lower thrust.
This does not help the cause. It makes SpaceX come off as reckless, irresponsible, and untrustworthy. It makes it less likely that the FAA is going to sign off on a launch from Boca Chica again, and certainly not in the next year or so.
There needs to be some new word for, like, the "humblebrag" on behalf of a personality-cult.
Like, that low-key way where when you pretend to be in complete denial about all the reasons why people might delight in the failures (perceived or otherwise) of a Nazi billionaire's ridiculous vanity projects and then come to the highly self-serving (well, cult-leader serving) conclusion that "they must just be impatient with the saviour of mankind for not being the saviour of mankind enough"
On the contrary, I find it more concerning that people are willing to discount the efforts of hundreds of people to progress human spaceflight just because the CEO is an asshole.
It is interesting that you would immediately default to a cult-of-personality retort when I never mentioned Elon in my original post. SpaceX is not just one man.
