3,418 karma · joined November 9, 2018
As for Kagi - they've earned a life-long ban from me for political reasons.
As I said a day or two ago, the open weights Chinese models may be good (hey might be better even) but let's not fool ourselves - the only reason the Chinese are pushing for open weights is because they were not the first people at the party and that's the only way they can fight against openai and anthropic. They are just as much of an enemy as both openai and anthropic - hopefully enough people realize that without supporting either. Just sit back and let them eat each-other. That's the only sensible thing to do.
As for China - I think their motivation is clear: if they could have pulled it off from the start, their business model would have been the same as MANGO's. Llamacpp opened up the floodgates and their best bet now is to use this and their resources to start pulling the rug from underneath MANGO. I'm not a fan of the "enemy of my enemy" philosophy - they can both be enemies but wile they are at each-other's throats (albeit not openly), I'll gladly sit back and grab some pop-corn.
That said, I firmly believe that if AI is to survive, the future HAS TO BE local or near-local. Having said that, statements such as
> Docker causes security holes:
> It creates a god-service that runs as root.
> It deliberately punches a hole in ufw firewalls.
Sorry, none of those are correct IF you know what you are doing. Though I will admit, seeing people that know what they are doing is increasingly uncommon.
Also there is nothing wrong with developing inside containers. If anything, that is arguably one of the biggest selling points for containers - environment(s) you can crash infinitely at no cost.
I still dislike opencode for a bunch of reasons - the assumption that llms are immune to screw ups, being one. As for the default behavior - using cloud by default - I didn't know that(I do not use any AI for direct coding tasks) and if that is the case, yeah, this is bad. Undeniably a horrible decision.
Heavy users: "this (works)|(doesn't work)|(is there)|(isn't there) and I have no idea why"
0 or near 0 usage: "this (works)|(doesn't work)|(is there)|(isn't there) and I know exactly why"
So to sum up: absolutely not. Heavy users drag the rest down when someone has to switch context and spend hours figuring out why something they "created" does not work, only to get aneurysm from atrocious code, dependency hell, practically 0 security and stupid decisions.
Languages have a lot of subtle, underlying characteristics which aren't obvious at first glance. I speak 3 languages(all European but from different families), 2 of which are sort of my first, the third one I learned in my early teenage years and ended up studying in that language in uni(that would be Spanish). As far as computer science, imo, Spanish was horrible (the available resources in Spanish were pretty scarce in the early 2010s) so what I did was to use the English documentation and resources, study that and translate it back into Spanish last second, along with learning the Spanish terminology. From an idiomatic standpoint, Spanish had one more weakness: generally questions are not structurally defined so in day to day conversations, the only difference between asking and answering a question relies entirely on your intonation(and that truly broke my brain). But it seems that the mixture of those languages in my head had some other side effects: how I think to myself/reason, as in I don't think in a single language but the one that's most suitable for the topic I'm thinking about. But if I am(or have been talking to someone) about a specific topic in one language I'll carry on in that language in my head.
English is a great language for much of that: it doesn't carry too many irregularities and can be both verbose and on point at the same time. Others carry a lot of additional information which is commonly irrelevant and can be inferred at-runtime: prefixes and suffixes and grammatical cases.
From a computational standpoint, Esperanto may be the perfect language because you could write all the rules of the language into a program and it would generate 100% grammatically correct sentences in all instances. Disclaimer: I've only explored very vaguely what Esperanto is and people that know it have told me it creates a fair bit of ambiguities.
However, I can only appreciate all of those are characteristics and properties now but otherwise I could have gone my whole life without knowing about any of them. For reference, just a few weeks ago I discovered that I have something called aphantasia (https://en.wikipedia.org/wiki/Aphantasia) and had I not stumbled upon the Wikipedia article, I could have gone my entire life without realizing that "Picture an apple in your head" is not a figure of speech(which is what I used to believe).
Community - it's generally supportive but can be pretty useless if you wanna take a single step out of the norm - you're on your own pretty much - you get a ton of shrugs and "idk".
There's another side to that as well: many people (contentiously or not) realized that when something is free, then you are the product. Now look at penai, anthropic, google, etc. Anyone that has basic GCSE level math skills can work out that their pricing does not cover their costs. Some people are in denial about it, some don't care and some truly believe that they are not the product cause they pay what is effectively a symbolic subscription. Or all three, but still, you are paying for something you don't own.
I don't come from a wealthy family and when I was a kid, all the software I used for making dumb games like flash, photoshop, etc were pirated. Same with music and movies. Eventually I switched over to Linux and open source projects. When I grew up and could finally afford those things, it only felt right to pay for a netflix subscription, spotify and whatnot. But due to the vile invasion in my personal space and the 0 guarantee that I'll have access to my favourite song the next morning, I got fed up and went back to self-hosting and pirating(to a degree). One of my best friends is a musician and I know that spotify is a big f-u to most artists since they have a winner-takes-all policy which makes me feel a lot less guilty. And frankly, if it is something I enjoy, I'll just head on over to the artist's website and buy a digital copy as a form of gratitude(even though I have often already downloaded the music): an album which I had very high hopes for dropped yesterday, I listened to it, liked it, downloaded it and bought a digital copy about an hour ago. Despite having it on my navidrome library since last night. At the end of the day, the artist will get a better compensation that way compared to what they'd get if I was listening to them on spotify, even on repeat.
So while the author has the right idea, sadly it's only part of the story.
Edit: After further inspection, I think I'm jumping ship before it's too late. And I'll look, see if there's a way to lend a hand or two when I have time!
1. It is fast and snappy. Nothing comes even close besides vim (and I don't mind going full time to it if I have to)
2. The ability to completely shut off and block any slop machine features from interfering with my workflow or leak code back to sloppenai, sloppus or any other self-installed-worst-security-practice-backdoor garbage.
Having said that, I hope they don't remove that ability in the future and enforce the "slop is so good man, you should try it" philosophy.
At least with hardware, people are actually making something and have to use their brains.
Don't threaten me with a good time(also unironically).