abhv

I don't understand the negativity here.

He is a top expert on a security topic. Running Android platform security gives him an opportunity to have incredible positive impact for many people---which he did for a decade.

People weigh trade-offs.

At the beginning, he may have had high ambitions to deploy interesting, research-forward ideas to Android; at this point, he has accomplished a lot of that. Maybe now, he is considering other factors.

Guessing that people are only money-driven or have made some decision because of threshold personal wealth is awful, especially if you do not know them.

Almost all academics I know (I am one also) are driven by personal curiosity, intellectual ambitions, a need to identify and solve problems, and a strong desire for positive contribution. I know Rene and believe this to be true of him.

How to Stop Worrying and Start Using AI

20+2 (conditional support) versus 7.

22/29 = 76% in some form of "yea"

That feels like "rough consensus"

(1) in this case, an identity issuer provides the source of truth identity information. Examples include state DMV, your passport (you can try "Id pass" in Google wallet), etc.

(2) One of the goals of this project was to layer ZK on top of current identity standards that DMVs already issue, so that gov orgs don't have to change what they currently do to support the strongest user privacy. One example format is called Mdoc.

(3) The user holds the identity information on their device only. No other copies. The user's device makes the zkp proof on-device. This was one of the major technical challenges.

(4) The relying party (eg a website) runs the zk verification algorithm on the proof that is produced by the device to ensure soundness.

(5) Yes, the user can use any compatible implementation to produce the proof. We have open-sourced our implementation and we have a spec for the proof format that others can also reimplement.

My colleague Amit made a simple video explanation about zkp with Wired. https://youtu.be/fOGdb1CTu5c?si=EyBQS92WyeduIpH-

That doesn't explain the way this scheme works, but it's a nice start.

-- ID passes in UK, Arkansas, Montana, Puerto Rico and West Virginia. -- ZK technology for age assurance.

successive keys on adjacent rows of your qwERTy keyboard

If you owe the bank $10m, the bank owns you. If you owe the bank $10b, you own the bank.

See also this thread from yesterday, https://news.ycombinator.com/item?id=37147914

"Thousands of scientists are cutting back from Twitter"

Essentially, a survey published in Nature about twitter usage among scientists dropping, in part, because of the phenomena mentioned in this New Yorker article.

Eg:

"Žiga Malek, an environmental scientist at the Free University of Amsterdam, mentioned in the survey that he had started seeing a lot of “strange” political far-right accounts espousing science denialism and racism in his feed. He has to block them constantly. “Twitter has always been not so nice let’s say, but it is a mess right now,” he said."

This is a case study for an undergrad statistics or responsible journalism class.

* for traffic see the small note "all values rebased to 100" they are likely hiding the significance of the increase

* for the app downloads graph: does the 30x gap say anything? are there seasonal reasons that can explain why every Jan1--Feb4 has more DLs than Feb4--Mar11 ? e.g., new phones?

i certainly don't do clickbait. no links to the original or anything. i thought the content was this observation. i'm trying to use whisper.cpp to get transcripts for the older ones.

i was skeptical. But these four probably do 100 meetings a week each, many about new deals with new founders. They hear a lot of smart ideas and the best of those are going to leak out when the videos are all about trying to sound smart.

Maybe never...but he had the signal 2 weeks ago in E118.

Stripe charges 2.9% + 30c per transaction.

From that, about 1.3% + 5c per tx go to interchange + assessment fees.

This leaves about 1.6% + 25c for the payment processor.

14.3 on 816b is about 1.7% which is consistent.

On 1T, that means about 17b in revenue.

Lets assume they have 7000 employees (i've seen 6000-8000 in searches). As rough estimates, these SFO-based SWEs + knowledge workers cost 1m/yr on average (which includes their total comp, insurance, federal + state taxes, and operating overheads amortized over all employees). So their cost of labor may be around 7-8b/year.

They may have other acquisition and marketing costs, but it means the co can feasibly be earning >8b before taxes, depreciation, amortization, etc.

That number could justify a 80b valuation.

If their current round cap is 55b, then my #s on costs are off, or the multiple has dropped to 6-7. Please debug.

Perfectly conceivable that Arvind is punking us all by using an Llm to write that essay. Very basic stuff for someone who usually has subtle points to make.

A Turing Award winner (keeping it anonymous) related a similar story to me about his phd qualifier exam (when I was taking mine). This winner is a theoretical computer scientist who skipped most of his computer hardware courses. Near the time of his graduation, he was informed that he needed to pass a hardware course because it was required for graduation.

He worried; his advisor begged the hw professor to give him an oral exam in place, because he was a promising candidate with a faculty job already lined up and incredible published results.

Unlike Wein, the hw professor was happy to oblige and created a basic oral exam, suspecting that the student was destined for greatness.

The student desperately tried to cram the hw textbook, but his heart was simply not into the subject. He did however, get the main point of the entire field.

This is when the funny part of the story begins. If you know this person, you know that he is remarkably clever, charming and convincing. He tells me that the hw professor had prepared a small number of questions, all beginning with something like, "explain how a XXX works in a YYY architecture", and that he really didn't know any details of XXX.

He started by copying a basic diagram of a processor onto the board. He kept his answers short, he began each with quizzical look indicating that the question was obvious, and all answers included a remark along the lines of "Obviously, its for performance!" In some cases, while facing the professor, he would point backwards to a spot between components in the diagram. He would slowly move his finger to different areas based on the look on the professor's face.

The hw professor was amused and obviously passed the candidate!

TLDR: don't let classes get in the way of your education.

I totally agree with you. The hutli experience is also very special for different reasons.

low season, long weekend, 1850chf.

We are a family of 5, and in Switzerland, that means 2 hotel rooms x 3 nights. Staying in Zumthor's cabin is certainly a splurge, but marginal utility of staying in a special place well exceeds marginal costs.

Also, if you are irrationally nostalgic like me, the renting experience felt like how airbnb felt in 2011...before all the bad experiences which have swamped it in the last 5yrs

+1 even more so, because I could see the small flaws and gaps that Zumthor may tolerate in a personal dwelling, but may not fly in large-scale projects.

Lesson for me: perfection isn't necessary for the experiences I want from exceptional architecture.

-- Indeed, we didn't want to support that hotel, which is why we stayed in Leis. I'll add that to my post.

-- Like everything in Switzerland, it was easy to reserve spots online. I think we went on a Monday. Therme Vals is a well-designed space, and the spa experience includes ~18 different zones and temperatures. Some of them are gimmicks, but nonetheless delightful play of space, sound, temperature and texture (stone).

Our family isn't into the pampering-spa experience, but I think all of my children can still describe that place 1yr after---worth it for me.

1-- First snow of the season occurred our first night. My memory is that the walls never felt cold. We started a fire using that granite cylinder fireplace. You pack it with about 12 logs, light it and then shut the door tightly. It burns the wood efficiently over several hours and then continues to radiate ten+ hours of heat into the house. Overall: very comfortable.

I think the beams were 6" of solid wood.

2-- Alpine woods are incredible. The table top in that picture was unfinished. The instructions of the house say that if you spill something, you just wipe it off with a damp rag.

I was terrified of eating fondue or jam-and-bread with my kids there, thinking we would mess up his table. Fact is: we did have spaghetti and spilled sauce, and the damp rag treatment didn't leave any stains.

Zumthor also seems to have a big family and there is evidence his grandchildren operate like normal kids in that place.

My son gave me a wooden box carved from alpine wood, and the maker intentionally kept it unfinished. The maker told me that the oils in the alpine wood provide a natural protection; +1 yr later and that box sits in our kitchen holding salt. It has developed a natural patina despite lots of mid-sautee grabs, opens, and pinches.

TLDR: like cedar, ipe, and other special woods, I think the wood used to build this house will last a long time in that natural environment.

We also visited Therme Vals; that was the point of our trip. While I love the majesty of stone Vals, which have these huge stone slabs that hang above your head with small gaps, his houses in Leis are the human-scale version of his emotional approach to space.

Hiking, skiing, exploring. Some people still raise goats, sheep, animals in the high mountain country. They make exquisite local cheese from the milk.

(1) Any theory that encourages people to suspect working directly on important local causes that are meaningful to them, but instead, seek out the highest paying jobs in order to donate to EA causes is just hiding a special kind of ponzi scheme ("get more people in to fund the good work of the current people").

(2) EA creates an excuse to act outrageously evil "for the good of the longterm"

I was intrigued. As of Monday AM, status.twitter.com was still serving a DigiCert cert for "*.twimg.com"

Marc may personally feel one way, and credit to him for writing that essay; his wife and family may have a different opinion, and in families, we "disagree but commit" all the time

The letter was signed first by Laura, who did not co-author the essay.

Marc may personally feel one way, and credit to him for writing that essay; his wife and family may have a different opinion, and in families, compromises are a norm.

The letter was signed first by Laura.

I am a professor in a CS department.

* Easy: ask your dept for a job or RAship to continue to support and build the app and grow its utility. We pay students all the time to build website infra for us in the cs dept. It is usually a ~20$/hr rate, though, so you won't make too much, but it will fund your improvements.

* If you built the app on your own laptop, and you've published the app to the app stores on your own account, you likely own all the rights. Good for you. If the app is hosted in anyway on university resources, then you should be careful if you want to grow this into a business.

* From my own experience, it is very difficult to sell to individual university departments. Much easier to get hired as a consultant. But "enterprise software" that is sold by a license may require the "university procurement" people to get involved. They will went some evidence you'll be around for 3-5yrs, etc...

* Look into how Piazza, Gradescope, Overleaf are doing with university-wide account sales. From what I know, those sales processes are slow and expensive, even though the apps bring enormous value to faculty (e.g., i end up paying for my own single accounts and expensing them against my research budget). Use your network to find someone at those companies to do more research.

Good luck!

btw, happy to try your app and give you feedback

gus_massa: Since you are likely an expert, could you recommend a resource that explains how you use the Lagrangian equation for the standard model [1] to actually compute a predicted value for the electron's g ?

An elementary resource that goes through basic steps for a computer scientist (non expert in QFT) would be a great. A simpler particle than electron is also ok, but I'd love to understand how you mess with that equation.

[1] http://nuclear.ucdavis.edu/~tgutierr/files/stmL1.html

"if modern cryptography is broken": this statement has many interpretations.

In the context of the OP paper, approximately solving the t-bounded Kolmogorov complexity (in a precise technical sense described in that paper) is akin to breaking one-way functions.

A method to breaking one-way functions would in fact break all of the cryptographic schemes (enc, signatures, prgs, hashing, zero-knowledge, mpc, bitcoin...) that rely on computational assumptions that we know. There is then no hope for doing things like we do on the internet today.

A secondary interpretation relates to breaking a specific widespread cryptosystem like ECDSA or Ed25519 (which can both be broken with suitably large generic circuit quantum computers). In this context, maybe some important things break, but in principle, we can rebuild them using lattice-based schemes or something else.