Being an IIT grad mainly proves how well you could prepare for its brutal entrance exam (<1% acceptance). At work, I've seen IIT grads fare no better than other people. Sure, there are some brilliant people among IIT grads but that's to be expected with a filter that selects the top 1% of any population.
A less known aspect of IITs in the past is their gaming of the GRE/US grad school application process, ranging from straight out cheating in the GRE (in the paper-pencil version, IIT students were given a single block of time for all sections rather than having time-limits for the sections, a blatant cheat made possible by their self-proctoring), to creating GRE question banks by collectively memorizing the test, to using the Australian time zone (ahead of India's) to find out the questions on the GRE. When applying to grad school in the US, IIT students would divvy up the schools among the graduating class so that no more than 1-2 would apply to top schools and claim to be in their class' top 10%, regardless of actual standing.
So, anytime I see an IIT grad at work, I'm not really impressed by that aspect. The fact that Satya isn't from an IIT is of no consequence and actually makes him more credible in my eyes.
It doesn't feel like google has been a good steward of these archives. They blended usenet archives and google groups together in a really annoying way, content has erratically gone missing (see the lwn link that hprotagonist posted), and have seemingly lost interest overall--the whole google groups product has seemed neglected for years. My guess is the only reason it hasn't been sun-setted is that someone high up needs to remember it to kill it.
Yahoo is taking the final steps to kill its groups product in a couple months. Google is usually 2-5 years behind Yahoo, so expect Google Groups to die before the end of 2025.
It really seems to me that archive.org, or some similar organisation, should petition Google to donate their Usenet archive, which they clearly don't care about, to them. Or possibly buy it off them.
It's a long shot since it would require Google to actually do something with a dataset they clearly don't care about, but they could get some good corporate PR points almost for free.
I agree. However, the IA team hasn't really excelled in showcasing their awesome collections. The UX isn't great, the performance of the wayback machine is objectively bad, etc.
There have been reports that some of those posts have been vanishing from Google Groups[0].
It's nice that this person/org is hosting these, but I'd feel more comfortable if archive.org (and perhaps another org or two) mirrored them as well. History shouldn't be dependent on some random person/org paying to host them. I thought things would be safe with Google and I was wrong. I'm definitely less confident that usenetarchives.com will be around forever.
I hear the term 'emergent property' bandied about in relation to the mind as if using it somehow explains anything. It says nothing more than mind exists, somehow, yet we have no clue about its nature.
Scientists and philosophers agreeing on something means nothing as they have agreed on utter bunk before. The short of it is that we know little about the mind and have no idea how to even start expanding on the little we know.
> Scientists and philosophers agreeing on something means nothing as they have agreed on utter bunk before.
That is an extremely cynical position to take; one could use it to dismiss anything, even things obviously true. For instance, "I don't believe the Sun is powered by fusion -- nobody has ever gone there to take a sample. Sure, they claim to have all sorts of indirect evidence, and there is 99.9999% consensus on it, but scientists have backed things before that were utter bunk."
> It says nothing more than mind exists,
It makes a much stronger claim than that. There are many people who believe that consciousness exists outside the mind, and the brain is a kind of consciousness receiver, akin to a radio receiver, that picks up the signal and relays it to the body. The claim of emergent behavior is an explicit rejection that that mystical explanation that is compelling to many people.
> The short of it is that we know little about the mind and have no idea how to even start expanding on the little we know.
This is entirely at odds with reality. Is it your position that brain researchers haven't learned anything over the past 10, 20, 30 years? Clearly they have, so obviously they do have ideas about how to expand on that knowledge.
"Emergent" is description of its nature. An alternative is "fundamental property". If you study an emergent property, you don't need to find its structure as a fundamental object, and this sets general direction of research.
Re. the bridge analogy in the article: we had a bridge being built behind our building at work and they had announced the opening date a year in advance. Sure, when the date came the bridge was opened exactly on time. I was walking with my colleague across it and asked him, "How come we can't ever predict when we'll be done?". His answer was deeply insightful: "They build the same bridge every time while we build a different one".
A long time ago in a philosophy of math class, our professor offered this problem: upon hatching chicks have to be separated by sex. Experienced poultry farmers know how to tell them apart from feel but it's not a process they can describe. The problem then was: how would they teach it to a new hire? The solution offered was that they have the rookie hold the chick in their hands and guess while the experienced farmer corrects them. This was a great example of how tacit knowledge is acquired.
Not sure about the Starship numbers but if its is $2m per Hubble mass, then launching 200 hubbles would be $800m, not $43m. This however is not about sending bricks into orbit. The ELT mirrors are segmented and need precise alignment with each other to micron precision. Deploying a segmented mirror in orbit has still not been done (the James Webb Telescope launches in 2021); the engineering involved adds orders of magnitudes of complexity over a single-piece primary mirror such as Hubble has. Add to this maintenance costs: a mirror the size of ELT in orbit needs a huge maintenance program behind it to work at all. The lifetime cost for such a mirror would be in tens of billions, if not hundreds. This is why such large telescopes keep getting built on Earth.
$2m is the cost on Starship's Wikipedia page, which implies 10 Hubble masses. Now, I am not talking about taking 216 mirrors and combining them into a single mirror pointed at a single camera, but actually 216 self-contained mirrors and cameras. That is, whether they're put closer together or further apart, they'll be interferometry based. What's important isn't how well-aligned they are, but how well they know how they're aligned.
I think the reason these things keep getting built on Earth is because until now no one has dared to dream of putting that much mass in orbit for that cheap.
Self-contained cameras are not good enough. You need to actually quantum-interfere the photons from one optical path with those from the other paths to get the effect of the larger resolution.
You can do this with optical fiber, a pixel at a time, for sufficiently slowly changing scenes. Each big mirror is then responsible for collecting enough photons so they can all interfere constructively and deliver a useful brightness level.
The output goes through a diffraction grating to get a spectrum for each pixel, so they need it pretty bright.
>> Is this what they did to bring us the high resolution picture of a black hole? If not could that method be used by these hypothetical hubbles?
That was using radio waves, something that can be measured in each telescope, saved as data, and then combined in a computer. At optical wavelengths we cannot record the data as accurately/quickly and so must instead use extraordinarily precise mirrors and fiber optics to join the light from each telescope in real time. These links must be accurate in distance to within a wavelength.
It sounds like the parent is referring to a specific type of interferometer, because yes, this is what they used on the EHT and it did not have any such entanglement or quantum effects.
Satellites need to be incredibly well built because they are ridiculously expensive to deliver. If we can deliver them for a fraction of the cost, it no longer makes sense to build them so expensively.
True! The way microcomputers supplanted mainframes.
If you launch cheaper satellites more often:
- you get to iterate so you can make improvements at a much faster pace
- there is less problems from a single failure
- more scale effects
You can enter a virtuous cycle.
There are limits! It can be that it's hard to make a useful satellite once you go below some certain size. And reliability suffers if you try to skimp on good practices.
On many scientific satellites, the instruments are one off anyway and are complex, hard and hand built with astronomical costs, so it might not save so much even if the launch and the bus were a lot cheaper.
Microcomputers came about because, at some point, we could make small computers with just enough processing power that they could be useful.
Mainframes didn't really go anywhere - they are still doing heavy lifting in large corporations, and modern mainframes employ every trick we use to make x86's fast and then a couple that would make x86's prohibitively expensive.
I'm not sure that holds true, at least for the big telescopes.
The lifetime cost of the hubble is about 10 Billion and the development cost of the James webb telescope is approaching 10 Billion. James Webb will launch on an Ariane 5 rocket, which costs about 140 million per launch. This is about 1-2% of the development costs.
I've had migraines for 40 years, and suffered much before being prescribed Sumatriptan. It relieves the migraine completely in about 30min. Doesn't work for my wife who also has migraines but an alternate formulation, Rizatriptan, does. My migraines have also evolved over the years from intense day long affairs to duller 3-day ones. Sometimes I need a subcutaneous triptan shot which relieves it in, I kid you not, 1 minute.
I tried preventative medications but the cure seemed worse than the disease. One, Topiramate, made me forget things and lose my train of thought. No thank you.
I also tried keto for a while and noticed that I hardly had any migraines while on it.
There are a lot of options available now. If you suffer from migraines, go talk to a neurologist with current knowledge, not something they learned 30 years ago.
Topiramate (topomax) is the antidote worse than the disease. It turned me into a zombie, literally living in my head but too dead to even open my mouth to say “hi” or “good morning” when I made eye contact with a stranger (yes, I’m from the Midwest and we do that here).
(Venlafaxine on the other hand is like a shock collar wrapped tightly around your brain waiting for you to miss a dose by even two hours, after which it’ll punish you for the next twenty four hours with electric shocks and spasms in your freaking brain. It’s a real pity it actually kind of worked :/ I still don’t know if I would ever do it again.)
A less known aspect of IITs in the past is their gaming of the GRE/US grad school application process, ranging from straight out cheating in the GRE (in the paper-pencil version, IIT students were given a single block of time for all sections rather than having time-limits for the sections, a blatant cheat made possible by their self-proctoring), to creating GRE question banks by collectively memorizing the test, to using the Australian time zone (ahead of India's) to find out the questions on the GRE. When applying to grad school in the US, IIT students would divvy up the schools among the graduating class so that no more than 1-2 would apply to top schools and claim to be in their class' top 10%, regardless of actual standing.
So, anytime I see an IIT grad at work, I'm not really impressed by that aspect. The fact that Satya isn't from an IIT is of no consequence and actually makes him more credible in my eyes.