This analogy makes no sense to me and honestly skews pretty elitist in vibe. iPhone is regularly used in professional videography now. Like, 28 Years Later was shot on iPhone. Indie filmmakers have been using iPhone to break into the industry for years.
the analogy would be that your LLM/agent has a pass at a Spielberg script and peppers his inbox with inane production notes. A system like that would be untenable for all involved.
I think the attitude frequently adopted by open source maintainers - comparing themselves to Spielberg - has been a major roadblock to anyone looking to contribute to open source projects for years.
Why are you looking to contribute to open source projects? If you have a fix or a new feature, you can share the diff in variety of ways. The maintainers are not obligated to review, discuss, and accept your changes.
I’m not entirely following you. I generally don’t contribute anymore, but in the past I’ve found a lot of maintainers are not actually looking for collaboration, rather free labor.
I certainly understand things are different nowadays, I’m talking pre-LLM proliferation.
> I’ve found a lot of maintainers are not actually looking for collaboration, rather free labor.
Do you think that maintainers lack domain expertise? A nice bug report is way more helpful than a random pull request. A patch, even when correct, can be counterproductive, if it conflicts with the roadmap and goal of the project.
The goal of open source is to give you freedom in maintaining your own version and extending it. Collaboration is not a requirement.
Agree that even prior to LLMs those projects weren't terribly welcoming as per Linus' famous email comments (chalk it up to cultural communication differences :) )
I don’t know if it’s just me, and these days I do understand it given the widespread adoption of LLMs, but I’ve always detested the idea that I need to reach out and have a conversation with the maintainer before opening a PR. Especially (mainly) when the PR is simply addressing an approved GH issue.
I’ve had so many perfectly acceptable PRs rejected over the years simply because they didn’t “fit the vision” of the maintainer, despite being +1’d by many members of the community or even other contributors. I don’t even mean to imply they were rude or anything, just uninterested in actually merging anything where they didn’t architect the changes themselves upfront.
On one hand I get it, you’ve spent so much time building something it’s fair to want to hold on tightly to that level of control, but to me it's just always felt antithetical to the entire idea of open source.
Makes me feel like I’m not contributing to a true open source project, just doing free labor for someone.
If you think filming is the only skill needed to make a film, may I suggest looking at the very long list of names that appears at the end of the film of which only a few actually do filming? Takes a lot to know what to film, and how to be good at using the tools you have.
Similar is true for a lot of software. Credit list on video games… I don't want to say it "mostly" isn't coders, but only because I've not done an exhaustive study. My guess is the top will either be QA or art.
Artists of all stripes (including audio, animation, cinematographers, lighting, environment, textures, etc,) including tech artists, designers, writers, musicians… the ratio of functionality to look-and-feel is dramatically different than in non-entertainment products, and the labor involved reflects that. It’s a real shame that some of the people that contribute most to what makes a game great are often the first to get dropped when people talk about how the game is made, (but most are perfectly happy to fly under the radar when a bunch of entitled kids start raging about the “lazy devs.” ;)
Exactly - it's just a tool. So, trying to make the argument that someone's work is less-than because they used a cheaper/more amateur tool versus the tool the well-funded professionals are using _is_ elitist. You recognize that, but the comment I replied to centered on the tool, not the finer points of professionalism.
But on that- whether folks have knowledge and taste, demonstrate responsibility for their impact, pay attention to their work quality, show up to the work environment with respect, etc. are all elements at the domain of human relations. This discussion is conflating how people use tools with how people work with each other. The tools don't matter here. I think we're sayin' the same thing.
> So, trying to make the argument that someone's work is less-than because they used a cheaper/more amateur tool versus the tool the well-funded professionals are using
No. Just the fact that they have a tool does not automatically make them a professional, doesn't automatically make them skillful, and doesn't automatically make their output worth something.
This is the meaning of "When they shoot a little artistic clip with their nice modern iPhone camera, it does not mean they get to insert it into a Hollywood movie."
Using Apple’s preferred practice of using no article before iPhone (ie. never “an iPhone” or “the iPhone” or even “iPhones”) makes you come off as a shill, by the way. It’s like if you unironically put a trademark symbol after it.
You can make art with a literally piece of shit, or a toilet if you want to be more traditional, at least in 1917.
You can't be a craftsperson without mastery of your domain and its tool.
You can be a artist without craftsmanship and vice versa.
You can also be popular without any or both of these.
There is a lot to entangle there but the point is that it depends on your goal. You can judge others based on your own value system but there goals might not be yours.
I think a better analogy would be commissioning an artist to create a painting. Yes you provided instructions and decided which style you preferred, and maybe pointed some corrections you wanted. And you can be proud of owning that specific, unique painting. No you didn't create anything.
"Demons are deceptive by nature, and typically speak with humans for a specific purpose, such as securing mercy or lowering vigilance. They treat language as a tool, using words without truly grasping their meaning. ..."
"Fable" is not the name I would have chosen for a product that has to argue for the fact of its own economic viability but it is the correct mood. Perhaps Anthropic is trolling Zitron.
I don't remember the exact number, but I read somewhere that when Jack Dorsey announced a 40% layoff and claimed it was being caused by AI, his personal net worth went up something like $2 billion. When we outsource judgment to what "the market" wants, this is the result.
"Their Head of Infrastructure, in particular, was young; maybe 30 years old and apparently one of the best AI buildout and energy experts in the country"
What's wrong with distillation? Wasn't GPT a distillation of the world's internet? That's how technology levels proceed, by recursively consuming the previous ones.
It's absolutely mind boggling to see claims of model distillation being theft, a class of attack, and all sorts of claims all the while Meta is in court for copyright violation, anthropic has had to settle a case with authors. With distillation "attacks" at least they paid API fees.
As you rightly point out, that's not a great argument.
2) China distills and is therefore possibly not that competent.
I think that makes sense. If they only catch up to the frontier through distillation then 1) Their model will never be as good as the model they are distilling from. 2) They will never reach the frontier - they need someone else to do it first.
You gotta start somewhere and you can start at page 1 or page 10 and that time, energy and cost you saved starting 9 pages later can be put into making whatever it is you're building better than the original.
The US, and every other country, is full of derivatives or straight up copies. No one is getting super mad at the generic cheerios at the grocery store. It's hypocrisy.
You're correct and I shouldn't have used the word competent. Perhaps "and is therefore not elite enough to be state of the art"?
I also have a soft spot for deepseek because they write such readable papers. I don't have a degree in anything but with a little work I can understand their papers - which I really appreciate.
But I still think my point stands - if you need distillation you won't be SOTA
Deepseek Flash V4 really was a "holy shit" moment and deserves the praise/hype it's been getting from users. I have a multi-tier subscription strategy I've maintained for the last year of:
1. $20-$30 plan from first Claude now Codex for "SOTA"
2. Gemini via the extra $10/mo or so from my Google One plan
3. a cheap fallback plan.
Together it gives me plenty of head room/model performance for $40ish/mo, plus letting me compare the various models over time.
Originally I'd been using the Z.AI plan (that I'm still grandfathered into for <1 yr) as my cheap plan but wasn't keeping up with the SOTA progress and is slow/limited now. So I subscribed to the Opencode Go plan and use Deepseek Flash V4 almost exclusively and it is insane how much usage I can get for $10/mo.
I did the math on my Flash usage vs. what I'm paying Opencode and I'm typically not even exceeding $10 in API costs! So it's actually sustainable not rugpull pricing at least for me. I can pound it with requests/agentic loops and have it running for 30 min doing whatever the fuck and check back and have spent literal pennies for what would have cost $30+ on my work's Github Copilot plan.
I know enterprise world works under different rules and isn't price sensitive in the same ways as an individual but I truly don't see how this is sustainable for the US AI giants in the long term to maintain like 25x+ markup for 1.25x performance benefit.
IMO it does help explain the recent emphasis on secret, scary "super models" like Mythos to muddy the waters for decision makers with hype and FOMO at at time when companies are beginning to seriously scrutinize their token spending for the first time.
Man, I decided to try DS with a healthy dose of skepticism.
I canceled ChatGPT because I would be on vacations. Codex was pretty great, but I thought "Let me put 10 bucks on Deepseek API and plug it into Claude Code".
I was completely blown away. I found it even better than Claude or Codex. And those 10 bucks? It lasted for more than a month.
Agreed. I’ve cancelled all plans except OpenCode Go. OpenRouter for API spend. Feels so nice to a) not feel like I have to code when I don’t want to just because I need to make my subscription worth the cost and b) know that this level of performance won’t be yanked away. Super pleased with DeepSeek V4 Pro.
I'm simply a postgrad from a no name university, but I'm not sure that the future frontier will come from the current approach from OpenAI and Anthropic. Distilling just seems like another avenue to collect useful data, like using books or scrapping the human-net, not necessarily copycat behavior. Chucking another 1000 TWh and scaling with already pillaged human-output is having diminishing results.
The next "frontier" as in, an order of magnitude higher model capability, might come from eschewing the bitter lesson and trying to be clever instead of pilling on GPUs. (I'm sure US labs also try clever new architectures) Maybe the plateau is permanent in the medium term. The frontier will be the exercise in taking the 10T monster models into something that can be run locally with minimal degradation.
That jibes a lot more with "AI is just new tech" attitude than the (genuine or otherwise) "we will build the Godhead in our image".
Those allegations reeks of projection and as far as I can tell in the case of DeepSeek - it is simply not true that their model is a distillation of a Western model.
Both. Both are good. Anyway this shows how full of shit Anthropic are - if Mythos was so advanced as they claim - distillation attacks just wouldn't work.
I haven’t worked on any web app in months, I don’t use LLMs, I update my Linux system once a month, and I increasingly feel I should just not do anything, not install or update any software and for the love of God, do not touch anything that’s shipped with npm.
Most of my userspace apps are in Flatpak sandboxes (yeah they are not great), but otherwise it feels like isolation and airgapping is the most sensible solution for now, and it’ll get increasingly worse unless the vibe coders somehow learn how to write robust software.
It’s like during the black plague: the (software) world has become dangerous, we have no way to contain it, it is unfeasible to remove yourself completely from the world, so you better pray really hard you don’t catch the bug and infect your peers. How’s that for a field we used to call software engineering or computer science?
My recommendation for them would be next time to create a profitable start-up or one that is valued at less than $1T.
reply