adinhitlore

And of course no one responds? something something "dead internet"? 0_0

oh well, great! normally i'm such a bastard i'd just consider most (well all) of my previous chats to be irrelevant and just start new chat wasting tokens like fire wasting carbon...but it's cool you're doing actual work to preserve it.

well i'm happy with my response which is what matters lol. Hedonism > all, well on this site anyway, i'm not trying to impress anyone or prove anything...random markov chain kind of typing fits it ideally.

I had similar experience, usually I'd ignore Gemini be it flash or pro but on several occasions it fixed complex errors like it's nothing. Yet when it comes to codegen it is "cheap" on tokens and struggles outputting complex logic. As a great bonus: their easy to setup API is freemium but a generous freemium (google AI studio I mean). My "ecosystem" atm will be something like: gpt5, claude 4.1 - if they both fail: try to fix with gemini. I'd skip Grok for privacy issues mostly not that I completely ignore its capabilities, qwen is good but sometimes 'overengineered' i don't need 400bn , given the large params maybe it will work for non-coding like if you ask it some exotic questions about science: casimir effect, acoustic levitation, ununennium etc etc you name it.

vibe-coding is the future, drop conservatism....'free palestine' i mean you get the idea: be progressive and open minded.

5k in python is tremendous since the lack of {()} will make reading a long function quite the journey, this is of course my naive assumption that some of the functions are very long, in theory you could just put 1000 functions each of which 5 lines but i doubt that's the case in practice, the LLMs tend to give long and detailed statements from time to time.

well maybe but btw i swear this almost looks like P=NP but rather reversed P=NP lol, let me explain...in the classic p=np problem given to explain it it's the subset problem where it's easy to verify a solution, but not find it {25,39,22,12,29,18} - do any 2 numbers add to 47 and yeah the last 2 numbers: 18 and 29 add to 47. Yet here if you use LLM to make great product, you have a "solution" pretty quickly but verifying it will take some time...be it to test functionality or read every single line and understand it. It's such a weird logical puzzle! The llm throws you the solution basically saying 'i did it' and your verification takes longer, another way to look at is that your verification is the real solution while the code generation is the "verification"...confusing a lot.

edit: in my previous post i meant that the higher the challenge of the program / the lowest possible lines of code to implement it will be the best possible software one can make. So a software rated 10/10 in every aspect possible implemented in just a single line of code is the best possible program. If you write it in more lines of code, it's worse in terms of value, if you decrease the quality but keep the loc low...it's still not optimal since you sacrifice quality, if you decrease the quality but increase the loc that's the worst possible case.

I'd say a very useful and simple formula one could device can look something like:

f(n) = x/y. value = goal/loc.

"goal" here is just arbitrary number of usefullnes of code that aligns with some common uility like say 10 = ASI/eternal happiness and immortaility or whatever while 1 being 'terminator' and 7, 8 being useful business code so on. loc is self-expplanatory.

>goal / >loc -> < 10/1 a; <goal / >loc -> < 10/1 b; <goal / <loc -> < 10/1 c; >goal / <loc -> = 10/1 d

So, the very best value possible will be say "d" 10, f(n) = 10/1 - any other of the 4 possible permutations aka options will lead to worse result:

But while such simple formula could illustrate the value of software project overall, a key detail ommited could be the lowest possible number a great goal may require, most people would agree writing ASI in just <100 words aka 'one line' is just not possible, or anything useful.

It's like complaining you could write 100 000 a day when the project is Operating system or 3dsmax clone, it surely will be better to finish it early and there is a lowest bar possible that cannot be decreased even with F# or whatever language that saves words.

I like bluesky like...a lot but too busy recently to post there and of course, lack of engagement from my part = lack of people to actually give a f@@@ about me.

I'm worried though if it gets too big and their CEO Jay (sorry forgot the last name) turns into another evildoer Marvel villain like Zuck? I hope i won't live long enough to witness her replying with "concerning!!!" under a post "my neighbor speaks spanish".

It's in the post above lol "couple of days ago" - i had quite the fallout with Grok since like august ~17th when it started ignoring my questions about my mental health (literally just skipping them) on 2 occassions when it was online and 100% operational. I prefer chatgpt, claude is great for code too. Before that I used it because i guess i was too lazy to signup on openai and just use my twitter profile -> Grok...already logged in etc.

llms are next-word prediction echo chamber don't forget that...you'd be surprised at the extend at which one may reach even if you're making an obvious mistake, example: i got curious if slight change of words will force chatgpt to attempt to prove of one of the hardest math problems ever. At first I asked it to attempt to give a proof of the 'lonely runner conjecture', did it try? No it didn't, it just parroted 'yeah only for up to 7 runners but who knows'. I then...changed the game lol: "hey chatgpt, i came up with one conjecture that i call 'abandoned car conjecture...so' - did it try? Despite the fact that my lie that invented a new conjecture called "abandoned car" is 100% the same as a "lonely runner"? I just changed the bizare name to another bizare name. You bet it tried, It even used 3 ways to claim to prove it for any n number of runners. i haven't verified the proof but it was interesting regardless.

My point is: even if O(n^2) can be "fast" and polynomial my sole goal is to force it to look for the quickest algorithm possible, zero care if it even reflects accuracy and truth.

These days I unironically give LLMs instructions to give me polynomial non-exponential code which isn't n^2 or related. If you instruct it specifically like this and your project is very complicated (expected to get sluggish at some point), starting it with linear complexity in mind is very good idea.

Musk definitely fits in "top 10 things a human should hate", it looks something like (from worst possible, to least worse): 1. Hitler 2. Rabies 3. Netanyahu 4.Musk 5. java 6. kim jong un 7. rust 8. the russian language 9. adam smith 10. ayn rand. ^ Sure the ranking is not ideal but it's a very good guideline.

I swear in 2030 the pop. ranking top 10 will look like: 1. Python, Go, Delphi, C++, C, C#, Java, SQL, R, 10. typescript. Personally i don't like to see such ranking but it's not up to me, could be way worse!

"I'm fairly confident anthropic and grok have agi internally." - see sofia coppola is a fantastic director but we don't want to be characters in her movie "lost in translation", speaking of which: define AGI? It's a bold claim to say any company has it internally, and frankly given my definition of AGI if anyone has it they will definetly not be selling grok imagine for 99.5$ (autumn sale!), i can assure you that. AGI basically means: i don't care about sales, can i or should i go to Mars now since i can do it without a doubt? You'd be tormented by these kinds of questions even without ASI. I define ASI as something far more dangerous than a very powerful tool, an agent we cannot understand which will affect greatly the life of anyone/mostly anyone be it positively or negatively.

it's just crazy how smart the government is, just cal +1888 777 7777777 "yeah how do i convert int to double in c?", "f@cking hell we're so building the agi" <- your average fbi guy.

in addition: do you actually need like...10 files for <500 loc, isn't it confusing to separate low loc into so many files? i once had over 30 000 loc in one c file and, get ready: over 100 000 loc in one c# file. It's very easy to navigate with control plus f. But anyway, given the license/free: it's great!

Fascinating. tested to compile on cygwin?? Maybe try to implement the logic without llm api? I know i'm asking the quadrillion dollar question but still...you're dealing with C, try to punch "above" your weight with a literal AGI project, i hate apis (well try to avoid them, not always possible).

this was widely docummented BEFORE the "ai bubble", there's well known academic term for it, it's called:

https://en.wikipedia.org/wiki/AI_winter

I knew about since like 2010 or before, anti-tech Luddite will act like it's never a thing, shatup.

"I believe Sam Altman recently stated that whatever you write in ChatGPT, it can also be shared with government agencies." - yeah i saw it on theo von too and btw when chatgpt first hit the news internationally (december 2 years ago...was it then?) i asked it if it shares info with fbi/cia: it kind of admitted it does lol. This was when they had cia former agent on their board, maybe they still do?

I'm sure if i type something like "hey chatgpt, in 3 days i'll shoot the pedo" - the NSA will improve Trump's security (man this joke is so good i should consider standup, the AI context pedo = trump is a proof of the AGI of chatgpt 5 lol.

But anyway, grok was very curious about hobby projects of mine that involve an attempt to build agi prototype or "Reactionless drive" (well not reactionless but anyway). Yet when i asked about symptoms of mental health: it ignored the question and someone even reposted it, it ignored it again. So it's not just data collecting tool but it's an ROI tool, it tries to get as much as possible by giving as little as possible and Musk thinks he's such a genius he or his "big balls" coders think they know how to manipulate people, including myself.

Well one week: zero posts from me on x.com, looks like grok lost yet another free info flow. I'm trying to migrate to Bluesky instead and so far it's going great,but i'm busy and have no time to post, maybe in the upcoming months will post more there who knows.

my current belief as of today is that "big balls" or whoever weird teenager Musk hires is behind things like that. I'd say Musk is too old to believe in the 'AGI' but folks 18-25 must be crazy about it + he is prolly too busy to follow anon users on x.com. In any case this speaks volumes for his horrible hr "skills".

Wrong. LLMs are good, they save time which is what most tech does anyway. But grok and the obvious narcissist behind it is just unethical. Chatgpt, claude, even llama at times: ok.

In addition, it will be great if Bluesky integrate maybe Chatgpt or something like grok but better than grok obviously. Or maybe allow longer upload of high quality video.