HNHacker News
TopNewBestAskShowJobs

betterunix2

3,500 karma · joined September 18, 2017

submissionscomments
betterunix2··on The foolishness of "natural language programming" (1978)
Not without some significant changes in how we turn natural language into programs. Yes, you can use an agent to do it, but the agent is going to rely on a compiler for the same reason human programs do: to check semantic correctness. In theory you could build an agent that outputs machine code, but the results would be pretty terrible.

I am not sure we actually want natural language to be compiled directly to machine code anyway. Those rich type systems that Haskell people get excited about are actually automatic proof checkers, similar to Lean, and having an agent check its output against such type checkers greatly improves the code generated by agents. If anything, we will probably see more widespread use of the lambda cube type systems (with extensions e.g. double negative elimination) as agents become a standard programming tool; human beings tend to find such type systems very tedious and difficult to work with, but agents are good at dealing with tedious tasks.

betterunix2··on The foolishness of "natural language programming" (1978)
The most obvious reason natural language is a poor choice for programming is ambiguity.

Time flies like an arrow.

I saw the parade coming through the window.

For small programs that basically duplicate something else with a few tweaks, sure, natural language can be fine. "The same thing as A, but with a red border, and add a new button to do X." Beyond that things start to get tricky.

Interactive programming has far more value than natural language programming, and agents also give you that. An ambiguous natural language spec. that leads to the wrong result is less problematic when you can interactively correct the problem; "No, that button should have been on the left hand side." It is the new version of REPL.

betterunix2··on Quad9 (DNS Resolver) to Exit France
Copyright was established for the benefit of publishers, not authors. In the 1600s publishers were licensed, as part of a censorship system; the point was to only give licenses to publishers who followed the censorship rules. It was quite profitable for the publishers because they faced far less competition than they otherwise would.

The censorship system was abolished but the publishers wanted to keep the licensing system, so they came up with a new idea: licensed monopolies to publish specific books.

betterunix2··on Quad9 to exit France
Is there some reason I should have a problem with it?
betterunix2··on Quad9 to exit France
It is actually worse than that. This amounts to punishing someone for looking up a phone number in a phone book that someone else published.
betterunix2··on Quad9 to exit France
Copyright has never existed for the benefit of authors, and anyone who authors a scientific paper knows this. Copyright has always benefited the publishing industry and was established at the request of publishers in the first place, after they lost the monopoly they had under the previous censorship regime (the "star chamber"). Authors receive at best a small fraction of the sales of their work and most authors never receive enough compensation to actually live on, hence the expression "don't quit your day job."

At this point copyright is an obsolete system. It was set up at a time when large-scale copying required an industrial operation, and that is simply not the reality of the world anymore. We do not need to "reform" copyright, we need to build a brand new system that is more reflective of modern technology. Maybe that new system will actually focus on compensating authors as there is no need for a publishing industry, or maybe we will just accept that authors are not guaranteed any payment for their work and will have to make money some other way.

betterunix2··on Meta ordered to pay $25B penalty in New Mexico Cambridge Analytica case
I am not sure what "relative to the homicide rate" is supposed to prove. The homicide rate is not the same as the homicide conviction rate, or put another way, perhaps we have a lower incarceration rate relative to our homicide rate compared to other countries because we are solving a smaller fraction of homicides compared to those countries.

(Also possible, but I suspect less likely, is that our homicide rate is higher but not in proportion to the number of murderers; maybe we have a bigger problem with serial killers than other countries, and thus have fewer murderers to incarcerate relative to the homicide rate.)

betterunix2··on Meta ordered to pay $25B penalty in New Mexico Cambridge Analytica case
It is pretty small for a company like Facebook, take a look at their quarterly revenue and the various credit facilities and sources of financing they have available. They could shrug off a one-time $25B fine.

You need another one or two digits before you can say it is "not a small fine."

betterunix2··on Meta ordered to pay $25B penalty in New Mexico Cambridge Analytica case
> prove it beyond a reasonable doubt

Except that proving it is rarely necessary because the system relies on plea deals and because the 6th amendment is not automatically enforced. There are places in the US where the criminal justice system would break down if every defendant refused a plea deal and demanded a speedy trial.

Our jails are full because we have unnecessarily long sentences and a prison system that is meant to be punitive instead of rehabilitative (and which winds up encouraging recidivism and compounding the problem further). Things have improved over the past 20 years but we have not done enough to address the problem.

betterunix2··on Meta ordered to pay $25B penalty in New Mexico Cambridge Analytica case
Sorry but it sounds like you think "limited liability" should only be understood to refer to corporate debts and bankruptcy proceedings and that "corporate personhood" should be understood as a legal fiction to make things simpler for bureaucrats.

That's crazy talk. Did you miss the part where corporations are actually people?

betterunix2··on Meta ordered to pay $25B penalty in New Mexico Cambridge Analytica case
The fines are not big enough if the companies can just absorb the blow like that. $25B is not large enough to be a problem for Facebook. Add two more digits and the executives would be running to their doomsday bunkers.
betterunix2··on Meta ordered to pay $25B penalty in New Mexico Cambridge Analytica case
Yes, our obsession with jail is nuts. Why do I only hear people say so in response to populist calls to throw corporate executives in prison for actual crimes?
betterunix2··on The No Kings Takedown Is Why You Shouldn't Rent Your Megaphone from Zuck
More likely these activists chose the most convenient platform for getting their message out. IG is popular and that makes it an obvious choice. Also keep in mind that activists will generally use what they are already familiar with, and Facebook runs three of the most popular platforms i.e. platforms that activists were probably already using before they became activists.

The only reason activists would switch to something else is if Facebook gives them a reason to do so. I doubt this one incident is enough of a reason.

betterunix2··on The No Kings Takedown Is Why You Shouldn't Rent Your Megaphone from Zuck
There is one good reason: abuse.

Think about how long it took to get email spam under control (and even so it remains a problem). Think about the ongoing problems with spam phone calls. Now pretend you are someone with no technical background, someone who only sees the UI and does not really understand what happens behind it -- you are not seeing spam on closed systems, you are seeing spam on open systems, what are you going to choose? Also keep in mind that there are worse forms of abuse than spam.

betterunix2··on The No Kings Takedown Is Why You Shouldn't Rent Your Megaphone from Zuck
The problem is not that people are lazy; the problem with open systems is that they are much harder to moderate and people want effective moderation. Usenet was open, decentralized, and wound up dying because there were too many problems with moderation and white supremacists etc. would show up and flood NGs with their nonsense.

There are also legal issues that kill decentralized/user-controlled systems. Copyright and child abuse are the most prominent, along with terrorism and drug trafficking. Who gets sued or prosecuted when the system is open, decentralized, and user-controlled? The users, and that's why peer-to-peer filesharing died, why so few people are willing to host a Tor exit node, etc. Before talking about banning centralized systems we need to talk about updating the legal system so that it does not force centralization, which is a much bigger project and much harder to get right.

betterunix2··on Eight Bits of Trust; eMudhra's CA keys cracked in milliseconds
Sure, if they were actually sampling 384 bits. Sounds like they sampled a lot less than that; maybe someone though "384 bits" meant "in the range [0, 383]" or whatever.
betterunix2··on Eight Bits of Trust; eMudhra's CA keys cracked in milliseconds
I would be shocked if this was anything other than a bug. A CA already has the power to issue MITM certificates, so what reason would a CA have to deliberately generate bad keys?

In all likelihood it is code that reads 1 byte from /dev/random, or reads 32 bytes but writes it into a 1 byte array, or something to that effect.

betterunix2··on Eight Bits of Trust; eMudhra's CA keys cracked in milliseconds
Yes, what matters is the size of the range, not the bounds on the range. Also keep in mind that in some cases private keys are not uniformly sampled from a range of integers; for example RSA keys, which are sampled from prime numbers.

More generally, it is common to sample a bit string and use that as a seed for a PRG, from which the private key is generated. So in that case the range is [0, 2^k] and what matters is the quality of the randomness. Unfortunately high-quality randomness is not always available, especially for embedded devices and rack-mounted "appliances."

betterunix2··on Academics have a duty to defend open knowledge systems like Wikipedia
Actually the most significant reason to not cite Wikipedia is the churn -- citations are supposed to give readers enough information to actually find the specific information that is being cited, and that should be possible even decades after a paper is published. Citations to academic journals are generally preferred because journal articles can, in principle, always be found because of the archives maintained by university libraries. Citations to Wikipedia articles would need to include the specific revision being cited and there is no guarantee that any revision of any article on Wikipedia will actually be available 50 years from now.
betterunix2··on Academics have a duty to defend open knowledge systems like Wikipedia
It depends on what you mean by "open."

Yes, the reviewers' comments and the authors' rebuttals are generally confidential, and there are good reasons for doing it that way. One significant reason is to allow the reviewers to make their comments freely, without worrying about how those comments might be perceived by the public or how the authors' reputations might be impacted; likewise with the authors' rebuttals. Another reason is to avoid tainting the review process in situations where a paper is rejected from one journal and re-submitted to another -- the second set of reviewers should not know that it was previously rejected, nor should they know why it was previously rejected.

On the other hand, the peer review process is "open" in the sense that there is no explicit credential required to be a reviewer, only expertise in the field (typically determined by publishing papers -- in principle someone with no formal education could publish enough papers to start getting invitations to be a reviewer). It is also "open" in the sense that anyone can submit a paper for review as long as it adheres to formatting (page layout, font size, etc.) and length requirements.

betterunix2··on Academics have a duty to defend open knowledge systems like Wikipedia
What are "collective facts?" How do you distinguish "collective facts" from "alternative facts?"

It is hard to imagine a worse attempt at establishing a collection of commonly accepted truths than Wikipedia. Scientific journals are overwhelmingly better at weeding out falsehoods and hoaxes, despite the history of successful efforts to dupe reviewers. University libraries are also better at collecting knowledge, and in fact the entire industry of journal publishing began at university libraries. That knowledge is not locked away from the masses, it is just less convenient to access than Wikipedia.

Even Wikipedia has had to accept that a certain amount of oversight and review is necessary. There are some topics that attract bad actors, paid shills, state-sponsored propagandists, etc., and Wikipedia had to introduce "protected pages" to deal with such things. Add a structured review process and more rigorous citation requirements and those "protected pages" would be a lightweight version of academic publishing.

betterunix2··on Let's Encrypt is adding two additional remote perspectives for domain validation
The problem with email encryption is not about CAs or verification of keys at all. The problem is key management and how encryption interferes with the common expectations that users have about email. In a corporate environment it is easier because the corporation itself can manage user keys and issue yubikeys etc., but for individual users there is no such support. It is also clunky to have to first receive someone's public key before you can send an encrypted message; for chat apps that is not a problem because chat is interactive and typically real-time, but with email there is no guarantee about interaction.

Those issues are partially addressed with "identity based encryption," but IBE always has a "backdoor" because there is a party that issues the users' private keys (and hence can always read a users' messages). On the other hand IBE does allow you to "just send" an email to someone because the public keys are just the "identities" (e.g. email addresses) combined with a single master public key. Key management is not as much of a problem because of how keys are generated. IBE does have some users, but as you said it is mostly business/organizational use.

betterunix2··on Intel deprecates SGX on Core series processors
"SGX is hard because it's trying to preserve the open nature of the platform"

Except that was an afterthought. Originally only whitelisted developers were allowed to use SGX at all, back when DRM was the only use-case they had in mind.

betterunix2··on Intel deprecates SGX on Core series processors
There is already a blacklisting/revocation system for leaked SGX keys -- which failed completely when it was put to the test, after researchers tried publishing some keys they extracted on Twitter. However, it depends on Intel becoming aware of a leaked key, which makes perfect sense for the original DRM use-case and makes no sense in the cloud/server/hosting/etc. use-case.
betterunix2··on Intel deprecates SGX on Core series processors
I'm taking bets on how long before the next generation is completely broken as well. It is only a matter of time, and the timeline is likely short.
betterunix2··on Intel deprecates SGX on Core series processors
Except that the security model on the server side is broken as there is no way for Intel to know that a key is compromised and thus revoke it; at least for the DRM use-case sharing cracked keys on forums is common. Why would an attacker ever share keys in the tenant/host case? Moreover, is there some reason to believe that Amazon, Google, or Microsoft would struggle to extract a key if they are indeed malicious? Is there a good reason to believe that Intel would never just give keys to certain government agencies when asked? If you are worried about a malicious host, SGX/etc. are at best a partial, very limited solution even if all you care about is integrity/attestation.

SGX and TEEs generally are and always were a DRM solution, with the server use-case mostly being an afterthought that the marketing teams pushed hard. They also create a fantastic forced-obsolescence program as they require active support on the part of chip makers throughout an application's lifecycle; Intel can arbitrarily deprecate otherwise functional CPUs by just not revoking compromised keys (and perhaps releasing a few into the wild just to force people to upgrade).

betterunix2··on BGP Implemented in Go
BGP is the de facto routing protocol of the Internet. The basic problem that it solves is computing and distributing routes between independently operated networks. It is actually a pretty ramshackle protocol (among other things it usually uses TCP, which can create some interesting problems [1]) but it is the only widely deployed protocol that can handle the scale of the Internet.

In theory BGP can be used within a single network, though it is not well-suited to such an application except in very large networks. Facebook is known to use BGP internally [2]. Usually an "interior gateway protocol" like OSPF or IS-IS is better for internal routes, with BGP then distributing those routes to the external networks from the gateway routers (i.e. those routers that interconnect the networks). BGP can also be used for other, more niche applications within a single network, like setting up VPNs (VPLS, VPNv4, etc.).

[1] https://blog.benjojo.co.uk/post/bgp-stuck-routes-tcp-zero-wi...

[2] https://research.facebook.com/publications/running-bgp-in-da...

betterunix2··on People don't want to run their own bank
Have you ever dealt with an estate, or a disputed estate? I have, and I can tell you that technical processes would probably not accomplish very much. I have seen people go to court over whether or not a person was mentally competent when they wrote their will. What is the technical process for resolving that? If anything, a technical solution would make things worse in cases where an elderly person really was taken advantage of when they were planning their estate by leaving less room for a dispute.

The fact of the matter is that we have had the ability to create a technical process for executing a will for a long time. Banks allow you to designate a beneficiary who receives your assets after you pass away and that is executed automatically. We do not have to allow people to dispute that, and cryptocurrency does not really add much over such a system -- but we do not want that system because some objections are legitimate and disputes need to be resolved.

betterunix2··on Ask HN: How do carriers throttle streaming to 480p resolution?
They do not know, but they also do not need to know. Technically they are not limiting resolution, they are limiting throughput on their network; the streaming service will reduce the resolution to avoid long lag/loading times for users.
betterunix2··on Ask HN: How do carriers throttle streaming to 480p resolution?
Yes in my experience.
← PreviousPage 3 of 34Next →