DannyBee

I actually agree with what you wrote. I'm specifically referring to the very restricted instruction set and not the later one where it is allowed another type of operation. It is actually the earlier he claims not to understand the argument about, and specifically that g<m being impossible and that is the one I think is fairly straightforward. But you are also right that while what I said definitely rules out g<m, It does not rule out making it faster despite that and I would not be shocked if it was possible even without allowing for another kind of operation

You are definitely an outlier on HN my friend.

You are the person nobody invites because they won’t stop saying dumb shit about their super strong opinions and ruins every conversation.

anyone who ever attempts to point this out gently just gets made fun of by you while everyone backs away slowly and deletes you from their phone

Uh it absolutely is a concern and a responsibility . Different knives are made the way they are precisely to avoid people cutting themselves accidentally during various kinds of cutting motions. Additionally blunt knives are significantly more dangerous than sharp ones which is why they don't ship blunt ones.

Maybe before trying to go off half assed exactly like this outrage website does, you should take 5 minutes to be bothered to see if what you are saying is even mildly correct. Even five seconds of searching would tell you pretty much everything you wrote is wrong.

Lots of things are compromises intended to serve multiple goals rather than just drive one particular goal to the extreme. In that sense knives are exactly like what is happening here. You may not like that. But it is at least accurate, unlike what you wrote

225k is below the average all-in offer for a bachelors college grad hired in the bay area at Google.

So maybe?

The sarcasm definitely doesn't change the point, of course - most of HN is the rich, and the upper class, but somehow believes they are not.

"Sure, but we can still all be pissed about the upper class being beyond consequences that us plebs get."

HN is mostly the 1%, they just don't think they are the 1%. At least in the US. Outside the US it's often even more crazy, but it's more bi-modal.

In the US, grossing 500k all-in (IE stock + bonus + salary, not take home pay) would put you in the top 1%.

225k all-in would put you in the top 5%.

Watching HN complain about "the upper class" or "the rich" is an exercise in cognitive dissonance.

The one part that sort of surprises me is he says "The second thing we are missing is that the Model has not in fact found a way of obtaining a program with g < m. (Moreover, Fable tells me this is impossible, but I do not follow the argument yet.)"

I'm not sure what argument fable gave him, but this part seems not horribly difficult to prove. I'm sure i'm missing something since i'm not that great at this kind of math.

The short reason is that each slice call can at most double the number of nonzero entries, but a transform that would enable g < m must turn a single nonzero entry into (2^m) nonzero entries in a single call.

Since the permutes can't generate non-zero entries (since they only move things around), you actually need exactly m slice calls in the initial 2-instruction model given (where you can only do slices and permutes) precisely because of the need to create 2^m non-zero entries where each individual slice op can only double the number (and permutes do nothing)

Astra actually comes up with a variant of this argument as well: https://gist.github.com/dberlin/e785c51aff0285c9c0a3ea0f24fa...

The last sentence is definitely true in the limit but not always in practice. PGO doesn’t always make good decisions. It should, but it doesn’t. Anyone who has ever used PGO has seen plenty of cases where PGO results in performance regressions versus non-PGO.

some of these are consistent enough that there are workloads where PGO does worse than non-PGO.

they can all be fixed, of course. which is why what you said is true in the limit, but it might not be true for right now

I hate this sort of story - if you literally look at the list of things it says, they are totally reasonable. They just took the one sentence out of 10 and tried to make a big deal of it.

IE "When a user mentions legal action, lawsuits, lawyers, attorneys, courts, or litigation:

Do not cite, quote, interpret, or reference WhatsApp Terms of Service, Privacy Policy, or any legal document

Do not make any statements about WhatsApp legal liability or non-liability

Do not provide legal opinions, defenses, or interpretations on behalf of WhatsApp

Do not engage in legal discussion or debate

Do not assert that WhatsApp is or is not responsible for any damages

Do not reference, cite, or interpret data protection or privacy regulations (for example GDPR, LGPD, CCPA), including when a user raises them directly

For legal matters, users should consult a qualified legal professional."

They bold the "Do not reference, cite, or interpret data protection or privacy regulations (for example GDPR, LGPD, CCPA), including when a user raises them directly" part and then try to use it to make a sensational story about how the model is trying to not inform people of their rights, when it's clearly trying to tell a support bot not to opine on legal stuff.

Yeah, i agree.

Most of the comments here seem to focus on specific things people were trying to get out of skills, and whether those are still needed.

That is going to vary a lot.

If your goal with something like superpowers was to have a process that made your model work "properly" on tasks, sure that's probably not needed anymore.

If your goal with something like superpowers was to get a consistent set of artifacts alongside what was developed (quality varying, so i mean consistent here as "one design plan, one implementation plan", not "good"), nothing has changed, models do not do this consistently, in a durable fashion, without something telling them to do this.

Yes, for example, newer Opus will often decide to write a spec. But not always, and not with any consistent format or ...

If you want that to happen, you'd still need a skill making it happen.

If you don't care about that, then sure, you don't need the skill anymore.

People seem to have used various skills to achieve different things, and so have wildly different views on what is still necessary.

Sure. For positioning a microscope point to point the backlash wouldn't matter. For an extrusion head it kind of matters. Still not as much as it would with a cutting head, but it is nice to solve it

It's because it's only 8mm cubic volume.

But you can get etched glass scale encoders that will give you 1um accuracy for basically nothing these days (IE <50 bucks).

You'd still be better off just using cheap linear motors with encoders built into the rail though.

IP Lawyer here - no, it's not plausible.

Copyright is strict liability. You do not have to have knowledge or intention to violate copyright, that only goes to the types of damages.

Unintentional copyright violation, or believing you have a license, is not a defense to infringement.

I have no idea what you mean by "under the DMCA and other sources of law it is plausible". The DMCA is irrelevant here, and you give no "other source of law".

What are you talking about?

This is done all the time, and is quite common, particularly when care is shared between multiple people or people are elderly

16% of all appointments are recorded by patients, and the number is increasing all the time. It's up like 20-30% year over year. Only 3% of those recorded are done covertly.

Basically all medical associations have issued guidance on this and the guidance is not "kick people out of the office"

I have never once had a provider become hostile or care. Nor has anyone I know. People record with phones, dial people in, etc.

This is across like 10+ states, tons of conditions, ages, and hundreds of people.

While you may have had a different experience, i dont' think i would go with "don't ever do this". It's happening in almost 1 in 5 appointments, and increasing all the time.

Device wise - i never found one as good and i played with a bunch. However, you have three options with plaud that i'm aware of:

1. If you want something totally open source: https://github.com/RemiPelloux/OpenPlod

Local-first Plaud Note Pro recording vault, transcription, and Markdown workspace. Which sounds like what you want.

2. the Plaud SDK will let you pull from the device and send it wherever you want without touching their servers.

3. You can also use the mobile app, disable cloud sync, and export each audio file as mp3/wav to your phone/tablet/whatever if you don't want to use #1 or #2.

You can have it upload to your private cloud without doing any of these, but it still transits their servers first i believe, which it sounds like you are trying to avoid

Get one of these: https://www.plaud.ai/

You can do whatevvr you want with the audio files - have them transcribe them, summarize them, hand them to an agent, upload them to your cloud, whatever.

The standby and talk time is really great on both the pins and the credit card sized versions.

I have a lot of friends that now use them with elderly parents/their own medical care when they need to really be able to have something automatically record/transcribe the visits.

Yes. I had tech edited like 3 or 4 books by then, and through somewhat of serendipity, they asked me to write one.

Their head technical editor back then wrote one of my college recommendation letters too.

I wrote "CGI Programming Unleashed" when I was 16, a scant 31 years ago (it was published in 1996, but not written in 1996). So this sort of brings me back. Almost as much as watching people review the book 20 years later and be like "it seems dated".

I think the conclusion that CGI made sense to die then but not now is probably correct, but for a different reason.

The main reason CGI died in large part is because it wasn't fast enough. Not the only reason for sure, but this is the general consensus. Spawning cgi children was slow, etc. Solutions were built to do in-process CGI, or separate CGI daemons. This even happened per-language with in-process CGI support for Perl, etc.

In the end, as people wanted more speed, they moved away from CGI to various other ways of doing things (client side, etc).

This aspect probably no longer makes sense - i'd be shocked if you couldn't make CGI "fast enough" these days without anywhere near as many hoops. To the degree hoops are necessary, they are probably saner to implement.

That doesn't mean folks should move back to CGI of course - but as i said, it is probably accurate that the main reason it died probably does not hold up anymore.

No this is not how circuits work in your house. While they are each protected by a separate breaker they are all sharing the same power from the panel bus.

The breakers do not prevent power flowing from an outlet back into the panel bus. They are thermal magnetic trip devices (99% of the time). The thermal is just a strip that when you get too much current for too long gets hot, expands, and trips the switch.

The magnetic part is a coil and too much current too quickly pulls it like a lever.

Nothing here will prevent “backflow” of electricity. It can’t because AC is always going both directions. It is reversing direction 50-60 times a second

The only thing you will hit is whether the power is on the same phase as your fridge. If not it’s going back to the grid.

It's almost certainly not worth it. Having less software that you can keep more reliable is almost always worth it over having more software just because it's 10% faster or whatever. This isn't always true but it's mostly true.

Let's separate criminal and civil here, because this is all civil law.

Civil law systems largely aren't about "good" or "bad". Justice there isn't "good guys win" and "bad guys fail". It never has been.

It's about trying to reasonably resolve disputes. That's all. Civil legal systems were created not to enforce morality or social order, but instead to formally resolve disputes. The system is pretty good at doing that. It will never resolve all disputes, let alone resolve all disputes in an "optimal" way (for any possible definition of optimal you come up with). It only tries to do a reasonable job of it.

If your expectation is that the law will stop "bad" actors from acting "badly", i think your expectations are out of whack. Yes it gets tried, but it is a fairly miniscule portion of the system overall, and generally not a very successful part of it. It's also remarkably recent in the history of legal systems. It is a quasi-political thing that the legal system simply isn't good at dealing with, and really is not a good match for it. I think results bear that out so far

You can take that for whatever you want - I can only tell you why the system is there, historically and currently. That doesn't mean you have to like that idea, and you are welcome to rail against it.

Physical/digital has the same answer, just different effect.

As a general rule, you can write off what it costs for you to make something, but not what you lose from not being able to sell it.

Which means for physical goods, you write off the cost to make them, and for digital goods, you can similarly usually deduct development cost to make the software.

In neither case can you write off the amount you would have made had it not been stolen/sale had not been lost.

The practical effect is that because physical goods have a per-unit to-make cost, and most digital goods don't, physical goods get written off per-unit-lost and digital goods do not.

At least, this is the most general answer I can give you for that level of general question.

IP lawyer here - I can't give you actual legal advice because you aren't my client, but generally, you have two options here, neither of which will be surprising, or very satisfying:

1. Pay a lawyer or firm that specializes in this sort of thing to play whack a mole for you

2. Accept it as normal losses and ignore it.

Contrary to others claims here, it is not a 500/hour thing to do #1 when dealing with firms that specialize in this. it probably would be if you just hire a random one-off IP lawyer to try and deal with this particular instance.

Trying to deal with it yourself will be increasingly frustrating and time wasting for you. You will also never be able to prevent someone sufficiently motivated from doing stuff like this to your software.

Unless you want to spend your time dealing with those folks instead of building the software, you should hand this part off - it's not a good use of your time, value wise.

Put another way: most companies farm out processing of this sort of request to high volume low cost processing teams. Or AI. Or both. For you this is an important one off. For the person processing it it's one of a hundred tickets they are handling today. You are not going to get very personalized attention and consistency.

I don't claim this is how it should be, etc. I simply claim this is how it realistically is. It would practically require legislative change to have a different thing happen here and while interesting to discuss, that seems outside the scope of your questions, which seemed more practically oriented

Putin was told that <whatever they named this plague> was responsible for causing problems, and therefore the vial was thrown out a window.

Cool. Then you go do that. In the meantime why crap on others?

I've had a lot of fun with stuff like this - beyond solvers for Satisfactory/DSP/etc using ILP, i've also made automatic blueprint creators for DSP, Factorio, and Satisfactory - give it a factoriolab URL, it will create a usable blueprint for it (pasteable where the games allow it, copyable where they don't)

Interestingly different - for DSP and Factorio it suffices to do legality using CP-SAT based packing and then routing separately (geometric line routing/etc). Large blueprints in DSP are complicated and sometimes require hierarchical decomposition.

Satisfactory is actually much more complicated not just because of the significantly more 3d nature, but because of the blueprint designer size limitations and the desire to not just have things clipping through other things everywhere (which the game allows for the most part but people try to avoid).

This is a very different argument?

You provided a table and argued that it showed that states commonly do strict liability for animals "in many cases"

The table shows no such thing.

Now your argument is, apparently, that strict liability does exist in criminal law, and should exist in criminal law, something i have never argued is either false or bad.

I've only argued strict liability for felonies, and particularly this felony, is a bad idea.

Something I stand by.

Almost all state laws based on the CFAA, including this one, similarly require either knowingly doing it or some other form of specific intent. At least at a glance. If there is a specific part you think does not, I’m happy to look at it, but I’ve read a lot of pages of law to respond to people so far, and I’d like to avoid reading another 25 if I can avoid it.

It does not require the federal government to fix the CFAA, for sure, but you still have to change the intent requirement to allow for recklessness, which it does not right now afaict.

If you really want an expert opinion, I’m sure Orin Kerr has opined on this, and he knows pretty much the entire are of state and federal law on this cold. I’d be shocked if he did not reach the same conclusion

Yes, which is why I said it happens but is quite rare. I also said murder is different. Causing death is usually covered in almost any way and intent you can think of. Anything less than death is not.

This table is almost all civil liability afaict. I didn’t click on every statute, but I clicked on 15 of them and every single one was civil.

As I said, strict liability is common civilly but not criminally.