Don't "no coders" find the concepts described in this article imdimitating?
The article states that whatever the article is trying to describe "Takes about ~20-30 mins. The cognitive load is high....". while their literal actual step of "Googling "ffmpeg combine static image and audio."" gives you the literal command you need to run from a known source (superuser.com sourced from ffmpeg wiki).
Anyone even slightly familiar with ffmpeg should be able to produce the same result in minutes. For someone who doesn't understand what ffmpeg is the article means absolutely nothing. How does a "no coder" understand what a "agent in a sandboxed container" is?
we took a basic example and described it. (will try adding a complex one)
we have our designer/intern in our minds who creates shorts, adds subtiles, crops them,and merges the audio generated. He is aware of ffmpeg and prefers using a SaaS UI on top of it.
However, we see him hanging out on chatgpt, or gemini all the time. He is literally the no coder we have in mind.
We just combined his type what you want + ffmpeg workflows.
Wouldn't that intern just use an NLE (be it Premiere, Davinci Resole etc) anyway? If you need to style subtitles and edit shorts and video content, you'll need a proper editor anyway.
I love how the proposed solution is to essentially gaslighting the model to think that it's an expert programmer and then specify and re-specify the prompt until the solution is essentially inefficient pseudocode. Now we are in a world where amateur coders still cannot code or can't learn from their mistakes while experts are essentially JIRA ticket outsourcing specialists.
Love to see it. Especially locking out security features like SSO behind enterprise-talk-to-sales paywalls is basically extortion and hopefully a practise that will go away someday… but as long as the fuck-you-pay-me mentality exists, I don’t think they will.
The real key takeaway here is that Microsoft could fix this of they wanted: they have near infinite resources, the best people and more heavily invested in open source than anyone else in the business, but still refuse to even comment on the situation.
Dropbox Paper was the best Notepad-like app in the market, only to Dropbox to completely stop developing it almost immediately and then eventually making it worse by making it Dropbox-backed and now killing the app. It's a shame really.
GPT-5 is just OpenAI getting started. Just wait and see what GPT-6 is capable of and imagine that GTP-6 is just OpenAI getting started: if GPT-6 was a high school student, GPT-7 is an expert with masters degree; but GPT-7 is OpenAI getting started
There’s always a distinct lack of the names in the posts like this. What was the library that was being changed to what? You say it had ”no good documentation”, but it clearly has some sort of documentation considering the LLM did such a good job on the rewrite. Do you understand the ”large library” now?
Before I used my own proprietary code for media encoding/decoding. I also tested a WASM port of ffmpeg for a while.
Mediabunny's documentation might be fine for some developers, but personally I prefer a reference where I have a list of all functions and their specifications.
Personally looking at the documentation I would say that "no good documentation" is highly misleading, because the documentation that it provides is incredibly detailed from quick starts to detailed explanations, offers a lot of examples and has very high quality typings with inline documentation. Not to mention the code itself is documented thoroughly. Sure it doesn't have an API reference, but you get that from the typings, that what I usually do - just check the imports first and go from there.
LLM's will definitely find big uses in spam. However, it's not the _only_ use.
1) the code that LLMs give you in response to a prompt may not actually work anywhere close to 90% of the time, but when they get 90% of the work done, that is still a clear win (if a human debugs it).
2) in cases where the benefit from successes is as much as the potential downside from failures (e.g. something that suggests possible improvements to your writing), then 90% success rate is great
3) in cases where the end recipient understands that the end product is not reliable, for example product reviews, then something that scans and summarizes a bunch of reviews is fine; people know that reviews aren't gospel
But, advocates of LLMs want to use them for what they most want, not for what LLMs are best at, and therein lies the problem, one which has been the root cause of every "AI winter" in the past.
The MRR archive is probably the single most important collection in the field of punk music and I am glad that it has found a (hopefully) permanently safe home. Would be amazing if Center for Popular Music would digitize the materials - with the green tape and all - and index it for the public.
There's so many things there that nobody has probably seen or heard in decades, not to mention letters, notes and other additions along side records and flyers.
Also it's a fuzzy feeling to imagine one of my recordings is now laying around in a box in MTSU, waiting for someone to discover it possibly decades after.
Support your local libraries and archives and all the librarians and archivists!
The article states that whatever the article is trying to describe "Takes about ~20-30 mins. The cognitive load is high....". while their literal actual step of "Googling "ffmpeg combine static image and audio."" gives you the literal command you need to run from a known source (superuser.com sourced from ffmpeg wiki).
Anyone even slightly familiar with ffmpeg should be able to produce the same result in minutes. For someone who doesn't understand what ffmpeg is the article means absolutely nothing. How does a "no coder" understand what a "agent in a sandboxed container" is?