I see this directly. I'm an engineer but part of that is doing non-engineer work like making powerpoints, whitepapers, and generic work that's necessary but that I would delegate to someone else if I were high enough on the totem pole.
AI gives me the ability to do that stuff faster so I can focus more on the 20% and the quality of what I'm doing, not the nuance of formatting a document.
AI makes up for a lot of crap tooling, but is often worse than good tooling. In this case "good tooling" would be a document template or formatting language that is easy/fast to use. What I have seen AI do so far if you try to one shot a PowerPoint deck is underwhelming to say the least. (To say more: it looks like a word cannon was used to splatter text on slides with random positions, colors, and fonts.)
Khm, ackshually... any documentation work is best not left to LLMs. They are excellent at generating many words with little value. Even sketching a doc and running through an LLM although a bit better than just a prompt "generate docs from this code" still suffers from similar problems. I prefer to write briefly, but doing it myself.
OTOH - formatting and other stuff not related to content, this is great to delegate, even though the results are often mediocre, but I really don't care about that at this point.
Just skip the powerpoint if you can't be bothered to create it. It's painfully obvious for anyone listening that the presenter didn't come up with or think through what they're talking about.
Not always an option if you're being required to make a PowerPoint. The content doesn't need to be AI generated. Just the organization of slides is a PITA and not a good use of my time.
If you're a small/FOSS project, it's not worth the time overhead to set up and maintain and troubleshoot.
If you're a large project, companies generally get more value paying for SaaS and letting devs do dev, rather than troubleshooting Jenkins for the billionth time.
The real value self-host brings is somewhere in the middle, where you need paid-level features with 0 fiat (and can pay in time instead).
I have a noninformed thought on this... I think a lot of these issues are actually because they're not on modern Cloud yet.
I think GitHub was rolling everything themselves when Microsoft bought them. Ever since, MS has been working to strategically shift to Azure but that's a difficult and careful process that takes time.
In the meantime, the AI boom hit and usage went through the roof, crippling aging non-scaled infrastructure that hasn't had a chance to be modernized yet.
A lot of issues might very well be "we broke something while migrating", but I think a lot are also simply capacity issues on legacy.
For someone the size of GitHub, the cloud is not infinite or infinitely scalable, and running their own datacenters is fine and completely normal and not at all "aging" or "crippled". Like the meme says, the cloud is just someone else's computer running in someone else's datacenter.
The traffic increases they cite as the reason are absolutely the reason. Those numbers are insane.
Oh, no, not that Europe. You were of course talking about the Europe with Spain (oh wait, La Liga has a cloudflare kill switch). You were of course talking about the Europe with Italy (oh wait, the Piracy Shield). You were of course talking about the Europe with Turkyie (oh wait, ...
I shouldn't need to - if it doesn't foundationally offend you that Spain has endorsed a private entity with the ability to routinely turn off part of the Internet for their own private convenience then fundamentally we have a different view about how this whole thing should work.
but now the definition of europe becomes europe but not UK, since it's convenient to maintain the argument rather than discuss the real underlying issue!
Also, vibe coding has really become a skill in itself. Does it allow the average non-programmer to go to production easily (and likely have bugs/flaws)? Yes. But as someone who used to be a programmer and just fell out of skill with all of it, I've found that what I do still have is a keen sniff test for good vs bad, and can guide the direction and architecture with proper planning.
Couldn't this be true in the other direction as well? Anecdotally I see developers putting a lot more scrutiny into vibe coded PRs, while AI code tends to be highly commented (by the AI) and potentially easier to read.
I've seen way more human comments of "I don't know what this does but if I remove it everything breaks" in systems.
I started on the Pro plan a week ago and was already contemplating jumping to Max. When I hit a limit yesterday I upgraded to Max and hit a limit again before seeing the news of the changed usage limit.
For what it's worth, everything seems fixed today.
The biggest benefit I see so far is the integration of ChatGPT into Siri to answer questions for me while I'm driving instead of giving me "I'm sorry, I can't show you that while driving"
I was looking forward to that as well, but Siri refuses to read ChatGPT's responses to me...which makes it substantially less useful. It's just as frustrating as being told "sorry, you'll have to unlock your phone if you want me to play something on youtube music". Let's hope they iterate, I guess?
I was going to second your observation: i wonder if the language models are just better than what they had been using, because Siri has become incredibly more accurate recently, in my experience.
You don't need to deal with the hassle of your own email server for this. Just buy a domain and use Fastmail, Protonmail, or any other service you trust.
AI gives me the ability to do that stuff faster so I can focus more on the 20% and the quality of what I'm doing, not the nuance of formatting a document.