My Utility+ podcast with Matthew Putman
Another Update (Sep. 15): Sorry for the long delay; new post coming soon! To tide you over—or just to distract you from the darkness figuratively and literally engulfing our civilization—here’s a Fortune article about today’s announcement by IBM of its plans for the next few years in superconducting quantum computing, with some remarks from yours truly.
Another Update (Sep. 8): A reader wrote to let me know about a fundraiser for Denys Smirnov, a 2015 IMO gold medalist from Ukraine who needs an expensive bone marrow transplant to survive Hodgkin’s lymphoma. I just donated and I hope you’ll consider it too!
Update (Sep. 5): Here’s another quantum computing podcast I did, “Dunc Tank” with Duncan Gammie. Enjoy!
Thanks so much to Shtetl-Optimized readers, so far we’ve raised $1,371 for the Biden-Harris campaign and $225 for the Lincoln Project, which I intend to match for $3,192 total. If you’d like to donate by tonight (Thursday night), there’s still $404 to go!
Meanwhile, a mere three days after declaring my “new motto,” I’ve come up with a new new motto for this blog, hopefully a more cheerful one:
When civilization seems on the brink of collapse, sometimes there’s nothing left to talk about but maximal separations between randomized and quantum query complexity.
On that note, please enjoy my new one-hour podcast on Spotify (if that link doesn’t work, try this one) with Matthew Putman of Utility+. Alas, my umming and ahhing were more frequent than I now aim for, but that’s partly compensated for by Matthew’s excellent decision to speed up the audio. This was an unusually wide-ranging interview, covering everything from SlateStarCodex to quantum gravity to interdisciplinary conferences to the challenges of teaching quantum computing to 7-year-olds. I hope you like it!
Follow
Comment #1 September 3rd, 2020 at 11:54 am
Civilization on the brink of collapse?!
How?!!
Comment #2 September 3rd, 2020 at 12:09 pm
Rahul #1: It isn’t obvious? Here’s what models say has an excellent probability to happen: the night of November 3, it looks like a clear Trump win. Trump declares victory. Then a flood of mail-in ballots, from the people who justifiably feared covid, shifts the outcome to Biden. Trump and millions of his armed supporters refuse to concede. Chaos results.
Comment #3 September 3rd, 2020 at 12:51 pm
I just donated $50 to the Biden-Harris campaign.
Comment #4 September 3rd, 2020 at 1:32 pm
Just donated $10 a week to the Biden-Harris campaign. 60 days left that’s 8 weeks or $80 total.
Comment #5 September 3rd, 2020 at 1:33 pm
Scott, with all respect, and I don’t think _you_ do it on purpose, but all this talk of Trump and his supporters not conceding a possible defeat is a terrible distraction. So far all evidence points at the other side finding it harder to concede defeats. Remember “not my president”, “don’t you ever normalize this”, “we won the popular vote”, “the Russians did it”, etc.
As an outsider with no stakes in this race, I have the impression that the current president has a decent chance of winning, for all the boring reasons: decent economic track record, almost no foreign policy blunders, incumbent advantage. The other side likely knows this and sets their sights further already. Portraying their candidate as a clear favorite (which is likely not true), they are setting up their voters for another bit emotional disappointment (“it was our win and they stole it from us”), to sustain the current “end-of-the-world” rhetoric and to keep people politically active and hysterized for the next years.
Comment #6 September 3rd, 2020 at 1:39 pm
I just donated $50 to the Biden/Harris campaign. Thanks for giving me the impetus I needed to actually act on my beliefs. Thanks even more so for matching donations!
Comment #7 September 3rd, 2020 at 1:41 pm
L #5: This is what’s in the news today. You and I might as well be inhabiting separate universes.
Meanwhile, doesn’t anyone have questions or comments about the podcast? 🙂
Comment #8 September 3rd, 2020 at 1:50 pm
Awesome, can’t wait to listen to this! And wow, that’s an impressive amount of money to donate. I’ll toss in another $28 to bring my contribution to a nice even binary 10000000
L #5:
One of the big mistakes that sane people on your side make is a failure to appreciate two facts:
1.) The right-wing coalition has been taken over by its most extreme elements; and
2.) the left-wing coalition hasn’t.
Yes, there were people saying #NotMyPresident. But Hillary Clinton and the Democratic establishment accepted the results of the 2016 election and conceded defeat.
You know who DIDN’T accept the results of the election?
Donald Trump. Despite the fact that he won, he never stopped insisted that the election was fraudulent and that 3 million illegal immigrants voted and that he actually won the popular vote too.
I’m pretty confident that no matter how the vote breaks down, Trump will say the election is fraudulent again. Whether he actually wins or loses.
The only question is, if he loses, will he get away with it?
Comment #9 September 3rd, 2020 at 1:53 pm
Sorry, I’ve only just started, I’m sure I’ll have questions soon; sorry if I contributed to derailing this discussion away from your new new motto… I’ll avoid any politics talk in this comment section from here on out
Comment #10 September 3rd, 2020 at 2:08 pm
@John Michael:
It would be too easy to respond that no, the left has been taken over by its own extremists for a very long time, pointing out the most obvious excesses.
But a more cynical truth is that both sides are equally pandering to their own extremes, consisting mostly of the younger, more energetic crowd which is more easily activated and manipulable for political gains. while the establishment is holding on to the power and status quo almost indistinguishably in substance between the two sides.
Scott, sorry, we’ll be asking about Rampart next 🙂
(obscure reference to Woody Harrelson’s AMA where he kept telling people to only ask about the movie)
Comment #11 September 3rd, 2020 at 2:22 pm
At this point, isn’t it better to disallow mail in voting and encourage in-person voting, to prevent the potential chaos from the mail in voting swinging the result and causing a dispute about who really won?
Comment #12 September 3rd, 2020 at 3:34 pm
Not to be the neurotic proofreader guy, but (as of Thurs 2020-09-03 4:33pm EDT), the final line of your motto has a word-substitution typo of “something” for “sometimes.”
That is, it presently reads:
Comment #13 September 3rd, 2020 at 3:54 pm
I. #10,
The big issue is the middle against the GOP, which has absolutely been taken over by right wing nuts. The Left might also be taken over by nuts, but it is a minor player, so no one much cares.
Comment #14 September 3rd, 2020 at 7:43 pm
L #10,
It’s amazing to me that anyone is still indulging in this “both sides are equally pandering to their own extremes” talk. The reasonable way to decide this is not to look at whatever outrageous single examples of any political view you can find, it is to look at the presidential candidates themselves, who each represent many millions of votes and are by definition the consensus of their political factions. One of these is actively fomenting violence and the other is explicitly denouncing it, constantly, from any side or origin. But don’t take my word for it- compare some recent posts on Biden and Trump’s Twitter feeds and try to tell me with a straight face that they are equally pandering to opposite extremes.
I strongly encourage you to go look for yourself, but just for posterity here are the most recent posts, as of my writing, by both:
Trump: “Nancy Pelosi says she got “set up” by a Beauty Parlor owner. Maybe the Beauty Parlor owner should be running the House of Representatives instead of Crazy Nancy?”
(a retweet of his own tweet from earlier)
Biden: “The President has the duty to set an example” (caption to a photo of him wearing a facemask)
Comment #15 September 3rd, 2020 at 9:34 pm
I donated $300 to Biden-Harris. Thanks Scott for matching all these donations.
Comment #16 September 3rd, 2020 at 9:59 pm
1) $100 more dollars for Biden-Harris.
2) I have a thought similar to Scott’s new new motto:
“In the first few days of the lockdown I wrote up the proof of
The large canonical Ramsey Theorem (or maybe it’s the canonical large Ramsey Theorem)
because, when people are rebuilding civilization, that theorem will be the first thing they
need for their efforts”
You can find my writeup here:
https://www.cs.umd.edu/users/gasarch/COURSES/858/S20/notes/canlarge.pdf
(I’ll probably make both a blog and an open-problems column out of it.)
Comment #17 September 4th, 2020 at 9:54 am
The link to the podcast is not loading on the website!
Comment #18 September 4th, 2020 at 10:27 am
Partisan #14
Right, and also one of those two tweets is funny as hell while the other isn’t.
(not surprising considering that one of them had two Jewish mentors – Roy Cohn and Howard Stern).
Comment #19 September 4th, 2020 at 11:37 am
I’m recommending Republican Voters Against Trump as another worthy place for contributions. They collect short videos by Republican voters against Trump in the hopes of giving support and inspiration to other Republicans and ex-Republicans who detest Trump.
I haven’t been keeping a spreadsheet, but lt looks like a substantial majority of them are voting for Biden.
Comment #20 September 4th, 2020 at 12:09 pm
Michael Ball #17: Wait, really? It just worked for me! (But thanks for posting the first comment about the podcast that was the ostensible occasion for this post… 🙂 )
Comment #21 September 4th, 2020 at 1:23 pm
This question is likely dumb, but anyway, why do extremely smart and successful nerds send their kids to public school? Isn’t it basically hell?
Comment #22 September 4th, 2020 at 2:31 pm
Whatever happened to the old motto (not sure from when), something like “If there’s one thing you should take away from this blog, it’s that a quantum computer doesn’t work by just trying all possibilities”?
When I originally found this blog, like 10-15 years ago, and knew a lot less about CS, this tagline was both super clear to me, and actually really informative.
Comment #23 September 4th, 2020 at 2:32 pm
nban #21: I’ve thought about that a lot, especially given that my time in public school was pretty hellish for me (in most respects, not all of them). But when we offered to move Lily to a private school, she said that she likes her school and likes her friends there. And while the math is trivial for her, the reading is actually not bad; before covid started the school was very successfully teaching her to read.
Comment #24 September 4th, 2020 at 2:35 pm
nban #14:
School is good for kids — I recommend it.
I don’t recall it being basically hell. It wasn’t even heck. Of course, that was the early 60’s, and maybe by now the devil has expanded hell.
Comment #25 September 4th, 2020 at 5:22 pm
For those of you who are skeptical about Trump not being willing to concede if he loses, in his mind the choice is between being the leader of the strongest country in the world and probably spending the rest of his life in prison. Even his latest urging to vote twice is a felony which his enemies would conceivably be willing to prosecute him for. He’ll do anything to avoid the second option. Regardless of how it happens, if he loses then he won’t concede in my opinion. I just treat that as a given.
Comment #26 September 4th, 2020 at 5:39 pm
I donated 100$ to Biden/Harris campaign.
Comment #27 September 4th, 2020 at 5:55 pm
I approve of your latest combination of mottos 🙂
In so many things, success is about finding the right balance
Comment #28 September 4th, 2020 at 8:36 pm
Unrelated to US politics and the ominous November.
Learning new material and computational efficiency: I have noticed, anecdotally, that people have a certain amount of aptitude in various areas (that part is uncontroversial, unless “blank slate” and “growth mindset” is your ideology), and learning new material becomes inefficient (exponentially difficult) once you are past a certain personal limit in that area. Thus one cannot become a major league player without the level of talent required, regardless of the effort put in. Similarly, if one has no aptitude for math, learning anything beyond, say, how to estimate a discount in a store, would be an insurmountable obstacle. If you have no aptitude for singing, you can probably learn to carry a tune if you start early enough, but you will never get to be as good as those who are natural at it without putting in much effort, no matter how hard you work.
This is somewhat complicated by the observation that people’s aptitudes can express themselves at different ages, though most probably show up before one is 20. And one might not know they have an aptitude for something if they never try it or connect the dots, or are told to do it. It is a common observation that a talent feels natural and not at all like anything special. That’s why you can see well meaning math grad students say “oh, just go through Spivak, it’s not that hard”, and the programmers say “Python is really easy, anyone can learn it”. We succumb to the typical mind fallacy and assume that what is easy for us must be easy for others, if only they really tried.
Comment #29 September 5th, 2020 at 12:07 pm
Scott #2: Remember to short S&P 500 in Nov! 🙂
Comment #30 September 5th, 2020 at 12:26 pm
Tamás #29: I’m not sure! If recent events are any indication, Earth could become a Venus-like cinder inhospitable even to bacterial life, and the S&P would merely briefly dip and then rally back to where it was… 😀
Comment #31 September 5th, 2020 at 12:55 pm
Scott #30 Yep, I don’t have a link to listen to the Audio 🙁
Comment #32 September 5th, 2020 at 1:20 pm
Does anyone else understand what’s going on here? Does one need an account on Spotify or something?
Comment #33 September 5th, 2020 at 2:46 pm
I don’t have a Spotify account.
The link works fine for me.
When I follow the link to nanotronics, it links to spotify, where I can play the podcast. The link is to “open.spotify.com,” which doesn’t require an account, let alone a paid account.
Here is a totally different link.
Comment #34 September 5th, 2020 at 2:59 pm
The link works fine for me. No login required.
Comment #35 September 5th, 2020 at 3:32 pm
Douglas Knight #33 Thanks, that link works! For some reason there was no link to follow when I accessed the page on my laptop, but there is one when using my iPad.
Comment #36 September 7th, 2020 at 1:28 pm
As for the podcast, is there a plain text version?
Comment #37 September 8th, 2020 at 6:01 pm
Scott,
in Dunc Tank, when it comes to whether QM is adding anything to the mystery of consciousness/free will/…, you mention that QM is still deterministic because the wave function evolution is deterministic, and even if we don’t know what particular state we’ll measure, we can still compute the probabilities.
But this kind of contradicts the previous points about QC – we can’t even really compute the probabilities in practice once the system has enough qbits.
You say that maybe the thing is that we can never measure the initial conditions, ok. But even if we had the initial conditions, we wouldn’t be able to compute the probability distribution, classically. And even if we had a QC powerful enough to represent the initial conditions, the QC doesn’t give us the probability distribution either. If we have a thousand qbits, and we don’t know anything about the way the system evolves, we just will never get an explicit description of the probability distribution, we wouldn’t even be able to store it for all 2^1000 values. With the QC, we can only sample it by running the QC over and over (does that even qualify as “computing” the distribution?), and then just count the outcomes, which for 2^1000 possibilities isn’t feasible either – unless we know we have a Gaussian or something really simple (Short gives the right answer or doesn’t), but then if we’re using the QC to simulate something a bit subtle, like the probability of a rare quantum tunneling effect or simulate some really rare and complex CERN particle interaction, it could still be very inefficient.
Comment #38 September 9th, 2020 at 8:13 am
Anon #36: Sorry, I don’t think so.
Comment #39 September 9th, 2020 at 8:18 am
fred #37: Why are practical difficulties in calculating something at all relevant to free will? Doesn’t the very fact that you could eventually calculate something, from input data that you have, mean that that thing is attached to the input data by iron chains of causation?
Comment #40 September 9th, 2020 at 9:50 am
Scott #39
Aah, thanks, I now see what you mean.
Comment #41 September 9th, 2020 at 10:52 am
Scott #39: Here is a possible relation between calculating something and free will. This is not at all worked out and may be inconsistent, but as a thought experiment, here goes. Suppose we imagine that only some bounded number of computations B can be performed by the physical universe. Maybe B is some very huge number, and maybe it is a bound in terms of classical logical gates, quantum gates, steps of a Turing machine, or whatever, but it is some bound. Suppose that, given data that you have observed about some agent, in principle it is possible for you to calculate the behavior of that agent, but the number of steps needed exceeds the bound B on the number of calculational steps you can perform. Does that agent then have free will? If you answered “yes”, then here is a refinement of the question: there are 100 agents. You can calculate the behavior of any given agent, but it will take you B/50 steps. So, you can calculate the behavior of any desired half of the agents (let us say there is no way to speed up the calculations in parallel) but cannot calculate the behavior of all of them. Do some of them have free will?
Comment #42 September 9th, 2020 at 12:57 pm
matt #41: ymmv, but I’d go with “no” and “no”…
Comment #43 September 9th, 2020 at 1:43 pm
I recently watched a video where Scott discusses his one an only foray into experimentation! He explains that he built a soap bubble computer – i.e. to prove that physics cannot solve the NP hard, minimum steiner tree problem and showed that as soon at the number of nodes climbs beyond a handful, you begin to see the the system settling into local minima (i.e. not solving the problem).
I wondered:
a) Does this tell us anything about the compute power of physics?
b) It seems (I have no evidence to support this!) that providing additional energy into the system (perhaps a good shake of the soap bubble computer) can help it progress towards/find the global minima – how is the energy used to help solve the problem that was at first too hard?
c) This is a weird one, does that mean that there are no naturally occurring entities (things?) that are NP hard?
I know this comment has no relation to the original post…… sorry!
Comment #44 September 9th, 2020 at 2:07 pm
Scott #52, interesting! I’m surprised it’s not “yes” and “yes”, in line with “free will = Knightian uncertainty”. Are you saying “FW => KU” holds but “KU => FW” may not hold, or you’re saying that “computational limits” does not implicate KU?
Comment #45 September 9th, 2020 at 2:28 pm
not matt #44: I’ve never suggested that “free will” (at least in the sense I’d understand it) could be grounded only in the computational hardness of predicting someone’s behavior—not in GIQTM and not anywhere else. I’ve suggested that it might be grounded in the in-principle inability to collect the relevant input data, which is completely different. A computation shackles its output to its input by chains of causality, while leaving the input itself freely changeable.
Comment #46 September 9th, 2020 at 4:31 pm
Scott #45,
Thanks for the clarification*… but I’m not sure I get it. “Find the missing input data that explains our present observable universe, assuming it evolves according to the Schrodinger equation”: is it impossible to reduce that to some valid decision problem? In other words: in what sense does “lacking the relevant data” is fundamentally different from “hard to find because you have to try an exponentially large number of possible input data before you got it”?
* maybe one source of my confusion was your (superb imho) point that Penrose’s ideas about consciousness would be harder to attack if it was based on computational hardness rather than uncomputability (QCSD I think -not sure). Actually I even thought GIQTM could be described as a deliberate attempt to right this wrong.
Comment #47 September 9th, 2020 at 8:27 pm
Scott 45: a remark that in the context of black holes and firewalls, there are some arguments to the effect that certain thought experiments that might seem to violate causality (in the physics sense, not in the freewill sense) or violate entanglement monogamy don’t actually violate them, because it would take too long computationally to verify any violation. Not saying you’re being inconsistent, just saying some philosophical arguments like to treat this question the opposite way.
Comment #48 September 9th, 2020 at 11:05 pm
The ‘Dunc Tank’ one is a fun one with all our old hoary philosophical favourites – consciousness, many-worlds and AI !
Scott, regarding AI timelines, I should tell you that OpenAI kindly gave me full access to the GPT-3 API, and the output from that is significantly better than the indirect public knock-offs like AIDungeon that you may have seen. So I asked GPT-3 to prove that P!=NP, and well… it fires off some interesting, impressive-sounding technical speel. I’m sending you 3 examples of it’s output for the thread ‘Is This Blog Redundant?’ – I doubt it’s ‘proofs’ make sense, but it sure seems to know a lot about computational complexity theory…
Comment #49 September 10th, 2020 at 1:53 am
not matt #46: For me, the two things are completely different; indeed, explaining the difference is a large part of what I was trying to do in GIQTM.
If the only reason why you don’t know a particular fact—like, say, whether chess is a win for White or Black or a draw—is insufficient computational power, that means that the fact is mathematically determined by facts that do know. And the central property of mathematical determination is that it doesn’t weaken, not even slightly, as the chain of inference gets longer! It merely becomes harder for human beings to follow the chain—but that’s a fact about us, not about the chain. This, of course, is directly related to the usual case against free will, based on the mechanistic nature of physical laws.
But if you don’t know a fact because of insufficient data about the state of the universe, the situation is totally different! For then it’s at least logically possible that the information you’re missing could “change under your nose”—i.e., that it could be influenced by causal factors that you haven’t yet accounted for. Indeed, this is still true even if our current understanding of physics would regard the data you’re missing as having been set in stone a long time ago, maybe even in the age of dinosaurs. For maybe a future understanding of physics will change that view! By contrast, no possible development in physics could ever cut the inferential chain that leads (for example) from the axioms of ZFC to Fermat’s Last Theorem, because the latter is just math. That’s the difference.
Comment #50 September 10th, 2020 at 2:04 am
matt #47: Indeed, my position is that computational complexity is relevant to some things, and is not relevant to other things. 🙂
See #49 for why I don’t think computational complexity can be relevant to free will (at least, not in the way most people want it to be).
In a similar way, if there were an actual contradiction in the laws of physics, then I don’t think that the computational hardness of uncovering the contradiction would be any sort of mitigating factor or defense.
Crucially, though, and contrary to a common misconception, this is not what Harlow and Hayden said about black hole firewalls! AMPS argued that, by decoding the Hawking radiation from an old black hole, you could create a firewall at the event horizon, and thereby engineer a breakdown of effective field theory … not of the ultimate quantum theory of gravity! Harlow and Hayden then argued that, while this is true, engineering the breakdown of effective field theory seems likely to require solving an exponentially hard computational problem. If you believe that, then effective field theory seems to be protected against low-energy violations by an armor of computational hardness—which of course is utterly fascinating if true. But it’s crucial to understand that, even if someone were to pierce that armor, they’d only break a particular effective theory (one that we already knew to break down at sufficiently high energies anyway), not physics itself!
Comment #51 September 10th, 2020 at 7:01 am
I personally don’t think that consciousness has anything to do with computation or “thinking”, after all our thoughts just appear as content of consciousness, just like any of all the other content – visual, audition, emotional, sense of self, sense of volition, sense of familiarity, language manipulation, dynamics of attention… and consciousness doesn’t give us any insight into them, they just bubble up in consciousness.
But I do think that consciousness is closely tied to the process of forming memories (and more clearly short term memories) and the nature of memories.
After all we can only remember what has been vividly perceived in the space of consciousness, no more no less.
Comment #52 September 10th, 2020 at 10:00 am
Scott #49, well, if explaining this difference is a large part of what you were trying to do, then I need to reread GIQTM with that in mind, and probably some time trying to digest that. Thanks for the cue!
fred #51, if we postulate that consciousness is tied to memory formation, then how should we explain that most people don’t recall their dreams, despite about 80% will recall they were dreaming if we awake them in REM sleep? Notice there’s about 5 time periods of REM sleep per night of sleep, which means we can be confident that the average person have at least one dream each night.
Comment #53 September 10th, 2020 at 11:18 am
not matt #52
I mentioned short term memories in particular.
When it comes to dreaming, we typically perfectly remember the dream memories when we wake up, at least for a couple of minutes, and we’re also conscious when dreaming (because memories are formed).
Long term memory loss can happen too of course, but when it comes to dreams we can write them down when we wake up, we then forget (or think we’ve forgotten) those dreams, but weeks later the memories of those dreams will reactive if we re-read our notes – it’s quite an interesting sensation, because they both feel new and familiar at once. It’s something some people can also experience during a seizure, they get overwhelmed by made-up content that feels familiar, but they also do know those are not actual memories (because nonsensical).
Whenever we remember a memory, it reactivates into consciousness all its content (if it’s an embarrassing moment, we’re feeling a good portion of the original shame or anger again), mixing it with the current real-time content. It’s very similar to just imagining something, except a real memory is also associated with a sense of familiarity. And the more we reminisce on memories again, the more they get burned in (by the same process tying consciousness and memories the first time around).
That feeling of familiarity is also very important (but is also just an appearance in consciousness) as it is the thing that prevents our brain from having to constantly “dig out” the entirety of our autobiographical memory at once every single moment of our life. We just remember enough of who we are to trigger that feeling of familiarity – we’re basically living our lives in a state of amnesia that we do not question (most of the time). Only when there is no memory to match the current real-time flow of content do we start questioning who we are and where/when we are (the brain keeps digging more memories to find an explanation for the lack of feeling of familiarity). For example, when traveling abroad, those first few seconds when we wake up for the first time in a new hotel room, and things aren’t matching our habitual memories (our room, our alarm clock sound, our body aches), there’s a temporary sense of alarm until our brain reactivates the memories that bring up that familiarity (“oh, yes, I’m currently traveling to Europe”).
Comment #54 September 10th, 2020 at 11:39 am
To add a bit – very short term memories are needed to build a coherent sense of continuity from one moment to the next. The capture of the latest new content appearing in consciousness corresponds to the formation of very fast memories, which are continually compared to previous memories from the last couple seconds or so (it’s like a series of rolling buffers).
If the comparison leads to a match within some small deltas, then the sense of familiarity emerges and the brain can shift its memory formation to more specific computational tasks, or shift its processing to unconscious well-learned patterns, in some sort of auto-pilot mode, aka being in the flow or in the zone.
If the comparison leads to a sudden big delta difference, our brain shifts its focus on the content that is specifically in mismatch (e.g. a loud horn just went off when we were reading a book), and very vivid memories of this event are now formed, and on the flip side of the memory formation/consciousness duality, that interesting content is now at the center of the space of consciousness.
Note that I do not equate consciousness with attention. Attention is the unconscious process that decides which changes in the internal state of the brain will now become committed to memory (and as a result appear in consciousness). It’s actually interesting to observe the process of attention by looking at a plain background and noting the shifts (or absence of shifts) in eye movements.
Comment #55 September 10th, 2020 at 2:40 pm
I donated 50 bucks to Biden’s campaign from your link^^
Comment #56 September 10th, 2020 at 2:44 pm
But I think they are counting potential voters(citizen)
Comment #57 September 11th, 2020 at 5:46 pm
Apparently, SlateStarCodex is going to start his own practice and move to a new blogging platform that protects its writers and offers monetization (a transition that should happen in the next few months). Hooray!
Comment #58 September 13th, 2020 at 1:21 am
mjgeddes #48: It’s interesting to see GPT-3 tackle a task that difficult. Unfortunately the “proofs” are full of non sequiturs. Someone should try to find a protocol so that GPT-3 produces genuine proofs of simple theorems in a well-known area (e.g. Euclidean geometry, number theory), and then build on that.
Comment #59 September 13th, 2020 at 2:36 pm
Thanks for the info on the fundraiser for the math olympiad winner. I am moved by his situation. I have donated and shared with others, such as UT Austin’s Saturday Morning Math Group.
Comment #60 September 14th, 2020 at 8:56 am
If you’re a centrist and turned off by Trump and Biden, there’s this
https://articlesofunity.org/
Comment #61 September 15th, 2020 at 12:48 am
Mea culpa, Scott, everything is wrong about this post; off topic, referencing two papers of extreme importance, “new information” published 12 and 14 Sept. on Zero Hedge (of all places).
Firstly, Vitaly Vanchurin posted on Arxiv submitted 4Aug 2020 (The world as a neural network), a non-quantum microscopic neural network capable of learning out of which relativity, quantum mechanics and presumably human classical consciousness emerge. Impressive but way over my head. Is it true that the way in which human generated computer algorithms enable neural network evolution is not fully understood? This may be of interest to you or you are already right across it or you are not interested.
Secondly, today’s date 14 Sept (your time) sees the following very extensive post on abovementioned website, “(Rogue Chinese virologist joins Twitter publishes “smoking gun” evidence COVID 19 created in lab.) Full paper included.
Comment #62 September 15th, 2020 at 6:11 am
Thanks for the heads-up, Filip #57!
Comment #63 September 15th, 2020 at 10:22 am
Hello Scott, do you feel like sharing your opinion on this paper if you already have one? I assume you got it in your mailbox already:
Fiber Bundle Codes: Breaking the N1/2polylog(N) Barrier for Quantum LDPC Codes
https://arxiv.org/pdf/2009.03921.pdf
Comment #64 September 15th, 2020 at 4:53 pm
Tommaso #63: It looks exciting! We’re supposed to have a talk about it in our quantum information group meeting at UT, so hopefully I’ll know more about it after that!