This has been my major concern, so much do that I'm going to be launching a tool to handle this specific task: agent conception and testing. There is so little visibility in the tools I've used that debug is just a game of whackamole.
Fascinating. That UI looks like Vercel. What node/graph library are you using? It looks like it would be really useful for some projects I'm working on.
Thoughtfulness is sometimes increased by touch time. I've seen various examples of this over time; teachers who must collate and calculate grades manually showed improved outcomes for their students, test techs who handle hardware becoming acutely aware of the many failure modes of the hardware, and so on.
Said another way: extra touch might mean more accountable thinking.
Higher touch: "I am responsbile for creating this report. It better be right"
Automated touch: "I sent you the report, it's right because it's automated"
Mistakes possible either way. But I like higher-touch in many situations.
Curious if you have links to examples you mention?
The teacher example was from one of those pop-psych books on being more efficient with one's time. I can't remember the title off the top of my head. Another example in the book applied the author's model of thinking to a plane crash in the Pacific. I'm sorry, man. It's been a long time.
1. get the LLM to generate a spec from the code (the spec is your new source code). Take your time here, make sure it is correct. Split out implementation details unless they are critical to the business logic.
2. Using the spec, begin crafting an ARCHITECTURE.md -- you put all your non-critical implementation details here. If you need to track requirements, require the LLM to reference the paragraph number(s) of the spec file as it writes the architectural spec. Take your time here as well (this is the second layer of your new source code).
3. Create a high level TODO from the architecture doc.
4. Create a mid-level TODO from the architecture doc (you'll know if this is necessary, perhaps chunk work up by function, feature, or architectural layer).
5. Create a granular level TODO where each task is very tight in scope (write this function, create a translator for this JSON de-serializer to create the app-level model for this object, etc.)
Then let a model go to town.
In the mean-time, you should also require testing. If you have the opportunity, you should have an LLM inspect, spec, and write comprehensive tests for the old code on the external API/interface. You'll be able to use this to test your new implementation to ensure all the corner cases and functionality are covered.
> The people developing standards are in the business of developing standards
Are they? Usually these standards consortiums are composed of the companies that develop products based on the standards, where their products gain value from having a standard (a Blu-ray player and a TV with no way to connect them together is worth less). Even if they couldn't gatekeep the standards they would still have developed them out of necessity.
There is no business developing standards. All the technical parts are written by engineers from the various companies wanting to implement the standard. All that's left for the standards association is to host a mailing list and potentially organise some in-person meetings. And hosting the resulting PDF doesn't exactly cost $4000 / download either.
These are standard business practices. They own IP. People want to use that IP. They say "pay us X to use our IP". People throw a tantrum because money. Instead, people want to capitalize on someone else's hard work for free.
I understand the ideas behind open source, and I think they are excellent. But I also understand that people and the businesses they operate want to make money.
> They own IP. People want to use that IP. They say "pay us X to use our IP".
The general premise of patents and copyrights is that you're going to do some development work and then you get an exclusive right that yields a competitive advantage.
Standards are different. The purpose of the standard is that Alice wants her output device to be compatible with everyone else's input device and Bob wants his input device to be compatible with everyone else's output device.
There is no competitive advantage to be had because the very premise is that everyone possible is going to implement it to maximize the network effect. And the entire industry has the incentive to want the standard to be good and put whatever good ideas they have into it because they're all stuck with it if it isn't. Meanwhile because of the network effect, everyone has to implement the standard because if they come up with their own thing -- even if it's better -- it wouldn't be compatible.
So all of the normal incentives from copyrights and patents are wrong. You can't gain a competitive advantage from it, companies have a preexisting incentive to make it good even without an exclusive right, and someone who doesn't want to pay doesn't have the option to try to do better on their own because of the network effect. And the network effect makes it an antitrust concern.
The result is that NDAs and royalties on standards are just a shakedown and the law shouldn't allow them.
> Standards are different. The purpose of the standard is that Alice wants her output device to be compatible with everyone else's input device and Bob wants his input device to be compatible with everyone else's output device.
I do think there's value and a lot of work in coming up with a standard that manufacturers agree on. It's a huge coordination problem, based on the idea of unlinking a standard's success with the success of, say, a hardware competitor. It's real work! And like.... HDMI is an invention, right? If that isn't then what is?
"we should have drivers for the hardware that relies on this tech" just feels like an obvious win to me though. The (short-term) ideal here is just the forum being like "yes it's good if HDMI 2.1 works on linux" and that being the end of the story
I don't have much love for things that mean that like VGA info online all being "we reverse engineered this!!!" so they're not my friends but I wouldn't succeed much at standards coordination
> I do think there's value and a lot of work in coming up with a standard that manufacturers agree on. It's a huge coordination problem, based on the idea of unlinking a standard's success with the success of, say, a hardware competitor. It's real work!
It's work they would be doing anyway because they all benefit from it, which is why it isn't a coordination problem. The known and effective coordination solution is a standards body. Everyone sends their representative in to hash out how the standard should work. They all have the incentive to do it because they all want a good standard to exist.
Moreover, the cost of developing the standard is a minor part of the total costs of being in the industry, so nobody has to worry about exactly proportioning a cost which is only a rounding error to begin with and the far larger problem is companies trying to force everyone else to license their patents by making them part of the standard, or using a standard-essential patent to impose NDAs etc.
> And like.... HDMI is an invention, right? If that isn't then what is?
It's not really a single invention, but that's not the point anyway.
Patenting something which is intrinsically necessary for interoperability is cheating, because the normal limit on what royalties or terms you can impose for using an invention is its value over the prior art or some alternative invention, whereas once it's required for interoperability you're now exceeding the value of what you actually invented by unjustly leveraging the value of interoperating with the overall system and network effect.
> It's work they would be doing anyway because they all benefit from it, which is why it isn't a coordination problem
HDMI: tech is shared between you and competitors, but you don't get to collect all the licensing fees for yourself
Some bespoke interface: you can make the bet that your tech is so good that you get to have control over it _and_ you get to license it out and collect all the fees
in the standards case, the standards body will still charge licensing fees but there's an idea that it's all fair play.
Apple had its lightning cable for its iPhones. It collaborated with a standards body for USB-C stuff. Why did it make different decisions there? Because there _are_ tradeoffs involved!
(See also Sony spending years churning through tech that it tried to unilaterally standardize)
> HDMI: tech is shared between you and competitors, but you don't get to collect all the licensing fees for yourself
> Some bespoke interface: you can make the bet that your tech is so good that you get to have control over it _and_ you get to license it out and collect all the fees
Except that these are alternatives to each other. If it's your bespoke thing then there are no licensing fees because nobody else is using it. Moreover, then nobody else is using it and then nobody wants your thing because it doesn't work with any of their other stuff.
Meanwhile it's not about whether something is a formal standard or not. The government simply shouldn't grant or enforce patents on interoperability interfaces, in the same way and for the same reason that it shouldn't be possible to enforce a copyright over an API.
That's definitely a thing that happened, but it's minimising so much other important work that it's misrepresenting the whole thing.
Do you know how much bandwidth six channels of uncompressed audio needs? Home theaters would be a HUGE hassle without a single cable doing all that work for you.
ADAT Lightpipe supports up to 8 audio channels at 48 kHz and 24 bits - all using standard off-the-shelf Toslink cables and transceivers. MADI can do significantly more.
Let's not pretend surround sound is a nearly-impossible problem only HDMI could possibly solve.
I... think you might be proving my point for me? The ability to have a single cable that can do video AND a bunch of audio channels at once is amazing for the average joe.
Don't get me wrong, I use optical in my setup at home & I'd love to have more studio & scientific gear just for the hell of it, but I'm the minority.
I'm not trying to defend the HDMI forum or the greedy arsehole giants behind them. The DRM inbuilt to HDMI and the prohibitive licensing of the filters (like atmos) is a dick move and means everything is way more expensive than it needs to be. Was just pointing out that parent's comment was reductive.
Correct! Now put that USB cable _inside_ a DVI cable, magically solve all the buffering problems that plagued the industry for several decades, slap on some DRM over the top, and you'll have HDMI 1.0 :-D
You just replied to someone who explained it was about the DRM, with 'nuh-uh."
Pivot much?
The rest of the capabilities were all being done for over a decade before HDMI came out, and quite well by some companies.
Sure, firewire was typically used for video plus two channels of audio, but it's a single twisted pair, and HDMI uses 4 high-speed twisted pair to transmit clock and data, plus another few pins for out-of-band signalling information.
Technically, HDMI is actually a huge failure. It wasn't until 2.1 that they started supporting compressed video.
Take a system, figure out where it has the highest possible bandwidth need, and then insert the communication cable at that point. Yeah, that's the ticket!
Before HDMI, some equipment did AV sync really well, and even after HDMI came out, some TVs still didn't do the A/V sync very well. The correct buffering for that has nothing to do with the cable, although it might seem like it because when the audio comes out of the TV, the circuits in there sure ought to be able to do delay matching.
The adoption of HDMI was, in fact, completely driven by HDCP.
I replied to someone who claimed HDMI's only purpose was DRM, which is wrong.
I haven't pivoted since the start of the thread. There simply was not a digital solution that could negotiate then stream video and AND 2+ channels of audio, all in one cable, that was supported by more than a small fraction of consumer and industry devices at once. Firewire (which you seem fixated on), for all it's many technical superiorities, had almost zero market with Windows users, or consumers in general. Set-top boxes used it in the US, but was uncommon outside of the US. Camcorders used it, but in 2002 when HDMI came out most people were still using film camcorders IIRC; digital only really became commonplace well after HDMI gained footholds.
I'm not saying the cable itself controlled clocks and handshakes, I'm conflating terms over the last couple of comments. I'm referring to HDMI, the cable, the protocol, and the connectors. And yes - HDCP had a huge part in how HDMI was pushed, which is both bad (introducing proprietary bullshit's never great) and good (larger adoption of standards that work well in the field).
Was HDMI perfect? FAR from it. But all these "there was this tech that did THIS facet better" is missing the point that I've stated a few times. It was a good solution to a number of small problems.
But to be fair, there is a standard that could have been used for digital video, SDI/HD-SDI, but the transceivers were expensive and it doesn't support any form of bi-directional handshake. There was already prosumer kit, mostly in the US, which had SD-SDI connections as an alternative to component. It didn't get popular in Europe mostly because of SCART.
I was once talking with someone who was very much involved in the process of standardising TV connectivity, a senior engineer at Gennum, and he said it wouldn't have been practical and SDI couldn't have been competitive with HDMI.
Oh, for sure. That and ADAT are great examples of tech that worked and worked well - and maybe even instrumental in HDMI's later adoption of optical tech in their cables.
I would say a fair compensation for the original work is fair, until certain threshold, after which they must invent new thing rather than continued benefit of an existing.
Say once they earned 400% of valuation or cost of invention or similar. there could be a system in place. But of course the people to regulate this has a natural bias, as they themselves would be hurt by it, most likely. So the vast majority, ie. the public is at an disadvantage, greed wins again.
Where does "invention" end and "standard" begin? If I come up with a new and better way to transmit video between devices, should I be allowed to charge for the right to interoperate with it? What if I don't want any interoperability and it's just for my own hardware? What if I just want certain select partners?
Yes. It seems pretty obviously true to me that there should be no legal right to prevent interoperability and no recourse against adversarial interoperability.
The right to say "Compatible with X" or similar where X is a brand should also be protected.
So I sit down and invent some wonderful new interconnect. It would be be a big advantage to put it into certain kinds of video equipment. I don't make any video equipment, so I license it to companies that do. Should this be impossible? New communications tech should only be created as trade secrets, by industry-wide consortia, or altruists?
This is getting close to arguing against IP as a general concept. Which I don't really object to very strongly, but presenting it as a special carveout for communication doesn't make sense to me.
> Instead, people want to capitalize on someone else's hard work for free.
This would only make sense if there _wasn't_ free video standards competing with HDMI. How is it that one group managed to do this for free yet the other group charges clearly exorbitant rates for a nearly equivalent product.
> They own IP.
That isn't nearly as valuable as they say it is. They only do this to prevent piracy and not to promote any useful technical standard.
> People want to use that IP.
People are _forced_ to because the same group practically gives away their technology under certain conditions so their connectors get added to nearly every extant device. I don't _want_ to use HDMI. I'm simply _forced_ to through market manipulation.
> want to make money.
Selling drugs would earn them more money. Why don't we tolerate that? It could be, under some torturous logic, be just another "standard business practice." In fact looking at our laws I see tons of "standard business practices" that are now flatly illegal.
The law is a tool. It can be changed. It should be changed. The citizens pay for 85% of it and while businesses only pay 7%. Why do their "standard practices" hold a candle to the "needs of the citizens."
It all stems from the companies behind the HDMI authority. It's basically all of the major AV device makers circa early 2000s. They wrote the spec and added it to all of their products. Displayport wasn't around just yet so HDMI just beat it to market. Since everyone needed an HDMI thing to go with their HDMI thing, everyone else jumped on the HDMI bandwagon. Although I'm really not sure how HDMI managed to get it's way into PCs. Displayport should have just cornered the entire market, it's very popular on business-class machines. I'm guessing it's because of HTPCs and people wanting to put big TVs on their PCs is what led to the adoption.
I think the HDMI connectors popped up at the same time screens switched from 16:10 (VESA compatible at the time) to 16:9 to be more cost effective for the manufacturers. But I’m not sure why. I looked at graphicscards and wondered why HDMI suddenly gained traction in the PC space even after the release of DisplayPort. I think this should never have happened.
Same thing applies to PCI. I can get USB specs for free from USB-IF. But the PCI and PCIe specs cost $4000 plus. Just so I can write my own PCI driver. Legally, I mean. Oh, there is external references, but what if I want the authoritative documentation? Should I have to pay thousands and thousands for access (!) to a standard that is ubiquitous in every sense of the word? There is, to me, a point at which ubiquity trumps any "IP rights" the standards org would have.
That’s true for earlier iterations, but definitely not for an actual HDMI 2.1 signal. I think you can still connect to a DVI-D monitor and the source will automatically downgrade, but I haven’t tried it in a very long time.
Yes, and we should say "no more making money from stupid things like secret technical standards"
Copyright and patent protection is afforded under the principle that said protections grant concurrent value to the people as is granted to the holder of the rights. Stuff like HDMI specs gatekeeping simply allows a select group of people to exploit licensing and seek rent. It doesn't provide any benefit to the people of the US whatsoever, and the fundamental principle by which the rights were granted is violated.
Copyright and patent protection is intended to incentivize and reward creativity, not to allow conglomerates of IP hoarders and patent trolls to exploit legal gotchas, to allow endless rent seeking, or empower megacorps to mass file endless vague patents so as to provide endless legal challenges to small competitors.
Copyright and patent law as currently implemented and practiced are fundamentally broken and far diverged from any principled, meaningful benefit to the people.
There are what, 2 publishers now? Five nines of commercially viable patents go to megacorps and universities? Seven nines of all music and media belong to conglomerates of one sort or another? Something like that.
I understand the intent of the original implementations of copyright, and maybe the laws even made sense for a few years, but either they were corrupt from the start, or they were so badly written that they never had a shot at achieving any sort of meaningful ROI for the price paid by the public.
How much money could PCI SIG possibly be making for the rightsholders with those fees? They’re not charged to members, they’re not per-seat (so each company only needs to pay once even if they have 100 engineers that need to read it), and they don’t include patent licenses for shipping actual hardware. Nobody’s business model is threatened even slightly by making the standards public.
And as we saw with AV1 vs H.265, the IP encumbrance of multiparty standards can create barriers that kill their adoption and the corresponding ability for rightsholders to make money off them. It looks like that family of encodings is going to die off, with basically zero interest from anybody in licensing H.266 when you’ll be able to build AV2 software and hardware for free.
Anti competitive "standard business practices" should be counteracted with good enough competition law that forbids them. As simple as that. So I totally agree with the above comment. They simply shouldn't be able to prevent open implementations.
"Hard work" is the worst way to make money at scale, so that argument rings more than just a little hollow, especially when defending access control based moneymaking.
When I'm in these situations, I try and put myself into the IP holder's shoes.
"if I spent the time, risk, effort, and money to develop the pre-eminent protocol and hardware used by most TV's in the world... would I want to give that work away for free?"
I think the answer is probably no for most people.
Because most of us are not the IP holder, they think this technology should just be free (as you stated earlier).
This lack of empathy and care for others (even IP holders) is largely why these draconian IP rules and contracts exist. It's why there are whole crazy NDAs around the HDMI spec. It's because every time someone is given even a slight look under the covers, they try and steal it, because it's worth a lot of money.
This is a nuanced variant of "this is why we can't have nice things" all over again.
The HDMI Forum isn't "most people", it's a non-profit run by some of the largest companies in the space that self describes this way.[1]
I think it is reasonable to complain when "someone" is being so hypocritical and arguably engaging in anti-competitive practices. How do the crazy NDAs in any way server the self stated mission of the forum?
Chartered as a nonprofit, mutual benefit corporation, the mission of the HDMI Forum is to:
Create and develop new versions of the HDMI Specification and the Compliance Test Specification, incorporating new and improved functionality
Encourage and promote the adoption and widespread use of its Specifications worldwide
Support an ecosystem of fully interoperable HDMI-enabled products
Provide an open and non-discriminatory licensing program with respect to its Specifications
> "if I spent the time, risk, effort, and money to develop the pre-eminent protocol and hardware used by most TV's in the world... would I want to give that work away for free?"
Only if you want people to use it. Developing a protocol is an investment in defining the direction a technology follows; the benefits are not best accrued by charging for access to the standard, but rather by leveraging the ability to direct the trend.
The alternative is that the licensing charge causes a bunch of stupid friction and prevents the standard from being truly universal.
EDIT: Implementing a standard is enough work, paying for the privilege to do so is often a non-starter.
> "if I spent the time, risk, effort, and money to develop the pre-eminent protocol and hardware used by most TV's in the world... would I want to give that work away for free?"
This is absolutely fine. But it should preclude them from becoming a public standard.
> This is absolutely fine. But it should preclude them from becoming a public standard.
Define "public standard". And how is HDMI one of them?
HDMI is a private bundle of IP that the license holders are free to give (or not give) to anyone. We're not talking about a statue by a government 'of the people' what should be public. No one is mandated by any government to implement it AFAICT: and even if it was, it would be up to the government to make sure they only reference publicly available documents in laws.
Devil's Advocate time. Would the result of that be better or worse quality public standards?
(I don't actually know what I think off the cuff - but it's the obvious follow on question to your statement and I don't think your statement can stand on it's own without a well argued counter)
It's a fine question. I think the onus is on public regulatory bodies responsible for the standards; if they aren't able to pay for the work to be published as an open standard, it wasn't worth the cost.
Standards also benefit the industry as a whole, and it's generally in the interests of the companies involves to participate in the standardisation process anyway. Charging for the description of them is just a cherry on top (compared to e.g. licensing any relevant patents), I don't believe it's at all required to incentivize a standardization process.
(this is of course looking at interoperation standards - regulatory bodies are going to be more concerned with e.g. safety standards)
The idea that you can “steal” knowledge and ideas is farcical. One reason why China is so good at iterating rapidly on technology is that this notion of intellectual “property” doesn’t really exist there. Any cool new invention is immediately iterated on by a hundred different makers.
And the reason to release a standard is to make your own products better. TVs would be awful if every manufacturer brought their own proprietary video connector to the table, and those manufacturers who grouped together to create a standard would accordingly dominate the market.
China quite literally and unambiguously stole trillions of dollars in IP, trade secrets, and data from research labs in the West by explicitly and systematically embedding spies, hacking, and blackmailing/threatening employees/students wherever economically beneficial information existed for nearly 20 years. And this is on top of the practice of CCP sanctioned theft from and screwing over of nearly every company that outsourced manufacturing there from 1990 onward. The fact that they finally have enough domestic knowledge to actually innovate as a result of that isn’t some testament to what you think it is.
If someone spends a billion dollars researching some new technology and you have someone exfiltrate the blueprints, improve on it slightly, and then undercut who you stole from in the market because you had no investment to recoup… you’re not some enlightened morally righteous free thinker. You’re just a parasite.
No. It's parasitic to horde human knowledge and seek rent for it.
If one is obsessed with the idea of maximizing profit above all other considerations including one's brotherhood to man, there's still other ways to do it that don't involve hoarding knowledge.
You are morally righteous to liberate human knowledge, it just annoys rent seekers. Honestly, annoying rent seekers should just be an immediate marker that whatever you're doing is probably ethical.
Weird to call it entitlement when the natural state of information is to be free. What's entitled is asking the government to enforce arbitrary restrictions on other people making use of some information that you somehow intangibly "own."
(Of course, it's fucked up that corporations can siphon up all this content and then try to twist the law to carve out an exception for their extra special use case. Information still isn't free unless you're an AI company, I guess.)
That's hoarding. The content is downloaded and then used to train proprietary models at no benefit to greater humanity. Thus some few corporations are robbing the commons and trying to rent it back to us.
There's a difference between "infringing IP[1]", "stealing IP", and whatever we should call AI training. And it turns out the worse the behavior gets, the less likely the law is going to recognize it as bad.
IP infringement is what we're used to talking about. This is when I go and give a stranger a copy of some music I don't own. Or when some sketchy ass guy resells IPTV services to an entire island in Greece or whatever. They're not saying it's their work, they're just refusing to pay the appropriate licensing fee for it. And sometimes we might even agree that a license fee shouldn't have to be paid. What the Linux video driver people want is for the HDMI people to say "yes, you can tell people how to light up this video card in such a way that it successfully negotiates a connection at HDMI 2.1 bitrates", which shouldn't even be infringement at all, but here we are.
What China does is wholesale IP theft. They don't just make their own version of someone else's thing, or just do industrial espionage, they actively make an attempt to deny the original creator of their own work. This can include things like forcing foreign entities to go through a JV, or playing games with trademark law to allow domestic companies to actually take legal ownership over foreign works. This is why a lot of American companies spent time and money carrying water for Xi Jinping, despite it going against everything they claimed to stand for.
AI training doesn't fit in either mold. It's more like rugpulling human labor by turning know-how and creativity into ownable capital distinct from that of traditional copyright and patents. Copyright gives you ownership over your own work, but says nothing about having your entire craft being automated away by a robot that can turn your work into legally distinct knockoffs of it[0]. So we have an entirely new form of enclosure of the commons, where if you ever do a thing, someone else can turn that thing into their own property that everyone else can pay to rent. Like, to be clear: AI is not Napster. AI is the opposite of Napster. AI is the apotheosis of "you will own nothing and be happy".
[0] The only way that copyright claims on AI even sort of fit into recognizable harms is the fact that at some point a Facebook engineer pointed LLaMA's crawler at a torrent site. In fact, I kinda hate how this is sort of saying "well actually fair use only applies if you bought the book first". Which is a problem, because the condition of sale can be "don't make a fair use of it", and the only way to avoid that was to pirate the work and then make your fair use.
[1] As Cory Doctorow said, paraphrasing: Intellectual property is the laws that allow you to dictate the conduct of your competitors.
The tracking app would be pretty simple. A shim in the shell that tracks the first argument on the command line and increments a counter. Then usage stats are pushed to the service every 24 hours (or something like that).
app usage / all tracked app usage --> donation to app.
Handling the payments portion would be... nightmarish?
With crypto it would be really easy. I'd be tempted to take cash on the frontend, convert to crypto, split, then reconvert to cash and make payouts.
How do we handle registering apps that you use? What about upstream dependencies? Example: you just spun up a new React project, do we target all deps on that project? Is this only for command line? What about cron jobs, or systemd? Or that systemd service script you copied from a gist somewhere?
I wonder if some weird kind of aggregation could go on.
At the end of the month, I'm told to donate my entire $10 to some obscure tool I've never even heard of.
The system did all of the tracking, across all users, and then figured out how to assign users like me to projects, such that I'm only donating once or twice ($6 to this project, and $4 to that project.)
The system I propose does all of the counting, all of the aggregating, matching intended donations to the recipients, and then encourages me to do my actual gifting...
Seems possible, but weird. Or maybe we need to set up a non-profit to do all this money laundering^W^W donating.
reply