v4dok

Thats not true. Ads are a race to the bottom (bottom being complete loss of privacy for optimizing sales). There is no one losing (not literally, but thats a s separate discussion) if the adds are "less" personalized. The industry will just settle to a less "optimized" outcome. Prime example: Newspapers were perfectly able to fund themselves with manually selling their inventory according to their reader groups before "the internet".

Making ADHD a joke is hurting millions of people who really have an issue. Yes ADHD medication is reducing apetite, but let me tell you, its not always nice, especially if you want to gain weight

>it turns out that low fat diets didn't make people less fat

This is not a "turns out". Anyone who believed that just having a "low fat" diet will make them leaner, didn't really know how losing fat works. Fat is a macronutrient (and quite useful one) just making it lower tells you absolutely nothing on its own.

I gave it a .txt file that i wanted to parse its data into .csv columns. It actually got most of it right but after spending hours prompting it, it never really got everything right.

Its worst nightmare was that there were some nested values in the .txt. What amazed me the most is that pasting the error code from colab into the chat, it recognized the mistake and generated a fix.

That being said, most of the times it would try to fix a logical mistake by overengineering a solution to an error, or it would identify the mistake correctly, explain in natural language what needs to be done (impressive), but then give me the code unchanged.

Long I wanted to tattoo this. My favorite (maybe false) story is that this was told to Roman generals when they came back from a victory. Kind of “be humble”. I don’t know if it’s true but having always in mind that death is there both humbles you and makes you appreciate things. Doesn’t make acceptance easier though and I think it’s normal.

I also think it goes perfect with mono no aware from Japan.

Know you will die but appreciate beauty in things even though they will finish/decay/die

Almost everyone I know having a big Iphone bought it because it was the most expensive one. No one does it because they like to hold a mini TV. I may be biased, but I have a wide sample.

Maybe the sophons are doing their work. Seems like a fundamentals problem to me.

Close. The first part is fair.

A more real-life example.

I am a pharma company and I want to execute a query on some hospital data. The hospital doesn't want to give me the data in plaintext but they are fine with me getting some aggregate insights from their data that are not PII.

Now lets assume I decide to do that using FHE. I can now compute my query on the encrypted hospital data and I never see the plaintext data.

What do we "win" in this scenario? We can do this computation wherever we want because no matter where the computation is done, the data will be encrypted, so no risk for the infra provider to see that data.

What we don't "automatically win" in this scenario? 1. Guarantees that indeed I am running an SQL query on that data and not something else along with it -> That is only possible to guarantee if the FHE software is properly audited (same with any software tbh, but easier with FHE and similar techs because of the integrity guarantees due to encryption). 2. Guarantees that the SQL query I made will not leak patient data in the end (through linking additional data, or diff attacks) (same with any other SQL query)

People who are deep into these technologies will say "yes of course" thats not an FHE problem. And that is true. But every FHE vendor I've seen blur that difference by not specifying what kind of attacks they protect against when they talk about "protecting privacy".

Heck, most of them they don't even talk about the attestation process and how their clients can make sure that they can trust the software running in encrypted form. Yes, these hold true for all software, but the point (for me) of encryption in-use is to make sure we hold software to a higher trust standard than today, not just replace a trusted party with another one.

You are right to be blunt. I didn't want to define the separation of input and output privacy in a comment. In retrospect maybe I should have but I can't edit anymore. It is however common HN practice to pick out the words that suit someone and construct an very specific argument based on these words, often missing the spirit of the comment.

Yes, when I wrote the comment I had in mind "output" privacy while FHE is dealing with "input" privacy. It is related to privacy, but not in the way most people think about it.

If you go to a random person and ask them about privacy they will not think about the threat model of a cloud provider leaking their data, but they will think of the thread model of a pharma company knowing exactly what drug they bought and when. That notion of privacy is not covered by FHE(alone). And even the first notion of privacy is covered only if the FHE program has a way to attest itself so you know that what you expect to run is indeed what is running.

You are right. It is related to privacy, but its not the whole story. Thats what I am trying to say. Running FHE will not magically solve the "data leakage" problem of your AI models, and I believe that the people/companies who don't make that distinction are misleading.

Yeah, and many more. But I've seen multiple people argue that using FHE will magically solve all their privacy problems and its far from true. FHE (and similar technologies) solve a piece of that "puzzle" and most providers somehow gloss it over.

Thats true. But it will still produce some data, and that data will be viewed by someone eventually who owns the key to decrypt it. FHE tells you nothing about what this product should be. It could as well be a full copy of the original data.

For example: I run an ML model using FHE on some data I shouldn't have access to in plaintext. The expected outcome of this workflow is a trained ML model on that data. FHE tells me nothing about the quality of this model. It could as well be an overfit model that spits out all the sensitive data.

I wouldn't hold my breath on this. Reducing it by x40 AND making sure that hardware is not leaky.

Depends on what do you mean by "person running the code"

If by person you mean the admin of the machine, then yes. If by person you mean the developer of the FHE-based application, then "maybe" If by person you mean the analyst who would in the end order an AI python model to be executed through the FHE-based software on a machine. Then no, that person will in the end get back human-readable results. Be that a model, or a table from an SQL DB running in FHE

"The question I would have is, if data synthesis, legal risk-based de-identification, differential privacy, and cryptographic tokenization protocols were insufficient, what technical improvement in actual accountability does FHE offer to data subjects, and given the size of the data sets this facilitates, what are the consequences of its failure modes?"

It doesn't. Because its not aiming at solving these problems. Encryption in-use is aiming to solve trusting hardware (and maybe code) you don't own. Privacy is a different (IMO more complex) problem.

I am still amazed by the amount of people that still mix security and encryption in-use with privacy. And truth is the whole privacy/security industry is doing nothing to change that.

Take this for example

"Valuable insights through AI (artificial intelligence), big data, and analytics can be extracted from data—even from multiple and different sources—all without exposing the data, secret decryption keys, or, if need be, the underlying evaluation code."

FHE gives you NO guarantee about the code that is running on the encrypted data. I can run a leaky AI model or a SELECT * on encrypted data and still get the output. What I can do (and that's assuming there is open-sourced, auditable code) is to make sure that anyone with hypervisor access on that machine cannot dump my data out during processing.

A very powerful concept for remote processing, supply chain security, and overall reducing trust; but completely unrelated to privacy.

Ultimately it's energy. And as long as the output is valued enough, we would put the energy there. It would be a terribly inefficient/poor world but it could be done. Unless we have so much energy available that the cost is negligible.

Definitely an entertaining read, so in that matter it "won" its place here. Now in regards to the actual content I am torn.

On one hand you could say he is bored, bitter, "old" dude who is tired from keeping up with the advancements and declares the end of progress.

On the other hand, you could argue that there is some genuine insight there. Internet is not an internal part of being a human. I even subscribe to the idea that it many ways it goes against being a human (social and physical detachment, desensitized emotions). And in the end of the day, no matter what is said and done on the internet, our existence is primarily physical.

Maybe the internet will not die, but maybe there is a regression to the mean. It feels like the excitement of the "digital self" is past its peak and people are slowly re-discovering that offline life is "cool again". Or if not "cool" a necessary plane to exist in order to effect changes. I think the Floyd example was on point; change happens on the street, not online. The stakes are simply not high enough there.

Important to note that we are talking about Turkey giving a casus belli on installation of AA systems in a foreign country. A foreign country that Turkey invaded some years prior to that, and still occupies 20% of the island. Pretty easy to draw parallels to today, and understanding why Cyprus, middle east and global south in general do not view the demonization of Russia as rational. These things were happening in EU neighbourhood for some time now. Cyprus is an EU member..

A bit late but hope you see it.

I mean that what we call the "competent part of Europe" is only competent in a myopic level. When it comes to their own local economy as if it is operating in vacuum. I argue that most of the foreign policy decisions of these members where made with short-term view of their own country economy, completely F'ing up their long term prospects by ignoring the potential and the dynamics of the periphery.

1. I meant declining in a more generic/long-term view 2. Independence is underpinned by a combination of military and economic power which adds up to enough power to take your own policy decisions. For now UKR has all the independence it wants to take any decision that is aligned with western goals. That (might) be aligned with their interests right now, giving the impression of independence, but it won't be always so. 3. I was trying to make conversation, your comment about my smoking probably tells more about you.

Its interesting because I would argue for the exact opposite of everything you say (almost). Which makes for a nice discussion.

I don't think the war will be over until US and EU decide to stop sending weapons to UKR. Now, in terms of UKR prevailing, no matter how you spin it, they lost 20% of their country and the majority of their fighting ability if not for external support. Doesn't look like prevailing to me. Basically any chance they had to be "independent" state are gone, either they are occupied or in heavy debt/reliance to the EU/US.

Yes future looks bleak for Russia indeed, but the same is the outlook for EU. A broken relationship with Russia never worked with Europe historically, and for good reasons. Also, Russia cannot be "helped" from the "west", its not "yet another" country. Whoever I know that has been there, can confirm that its a special place with dynamics that are not entirely understandable by western values.

I don't see China declining, the opposite actually. Yes COVID hit them (which for me screams "Cui bono" to the whole covid thing, but thats off topic) and they are doing bad demographically, but their struggles are not in vacuum. Europe is getting hammered and its even worse demographically, US is in a "esoteric mode" dealing with problems from within.

I would argue that its an engineering/technology problem at heart. If access to those things was easy and cheap politicians would have no/less incentive to hold it back. It's probably not the same if it's a limited and restricted resource.

If we had more people trying to optimize food production, water distribution, Healthcare scaling and management etc, those things would become materially better, fast.

But we are more concerned with optimizing ads so we can buy shit we don't need so..

Macbook Pro with M1 (is it M1pro I don't remember), but anyway, all of them have fans. I have yet to hear mine though, completely different from my previous intel MPB where fans were spinning on 80% almost constantly.

Your comment is flame. You really don't believe that competence is the only reason solar/wind didn't scale up right? If you remember, those countries were/are almost bunkrupt, since 2010 or so. Where would these massive investments come from? I mean the north didn't even want to move away from Russia, certainly they wouldn't care for funding solar in Greece, Spain or Italy.

The "Green new deal" however is a step in that direction. But IMO its theatrics. EU money to buy green German cars and green German infra with no plan on longevity or strategic change. Truth is, that if EU wanted energy independece, they could have it many years now. Be it nuclear or Mediterranean natural gas (the same gas now Israel is going to sell to EU).

Not that the respective governments are competent. They are not, thats a given. But the "competent" part of Europe is only so in a stupidly myopic level.

I think that was the point of the sarcasm

What about things like Trusted Execution Environments? I know that people love bashing them for being broken, but realistically speaking.. if you don't have malicious nation state level actor in your threat model(since legally there is no way to get the data, only by breaking it, and even then with a lot of asterisks), they could help a lot with moving things server side in an encrypted form.

The amount of those existing today would suggest "a lot".

I don't think modern big language models are conscious, mainly because they fail in absurd ways. But TBH, they don't need to. This "golden path" deployed properly etc could easily automate a lot of jobs tomorrow.

I find hard to believe that the ruling class in Europe thinks Pax Americana is over since Iraq. Especially considering how spineless they are in geopolitics terms. I(have) to believe that they still dream about americans coming to the rescue