Posts on X before Nov 15, 2024, 11:47:11 AM UTC
It's such a no-brainer that anytime I set up search for anything, I focus on making sure all setting names, as well as short descriptions (if available), are able to be fuzzy found.
Another fallback for me is usually ripgrep.
Search results should almost never be empty.
It's so interesting to see everyone talking about scaling thanks to Ilya's quote, but they don't seem to understand that there's different types of scaling. It just goes to show that a lot of people seem to rewrite high performing tweets in order to just boost for engagement rat… Less
It's so interesting to see everyone talking about scaling thanks to Ilya's quote, but they don't seem to understand that there's different types of scaling.
It just goes to show that a lot of people seem to rewrite high performing tweets in order to just boost for engagement rather than have something of substance.
It's so insane to me that so many people are reading Ilya's quote and the only thing that they're registering are the first and last parts. They completely remove the middle. All they see is "results from scaling up... have plateaued." Weird how no one is mentioning this is spe… Less
It's so insane to me that so many people are reading Ilya's quote and the only thing that they're registering are the first and last parts. They completely remove the middle.
All they see is "results from scaling up... have plateaued."
Weird how no one is mentioning this is specifically only regarding pre-training. And more importantly, it only mentions ONE of the phases of training an AI model that uses a vast amount of UNLABELED data.
I don't wanna say "I told you so", but I told you so. Quote: "Ilya Sutskever, co-founder of AI labs Safe Superintelligence (SSI) and OpenAI, told Reuters recently that results from scaling up pre-training - the phase of training an AI model that uses a vast amount of unlabeled d… Less
I don't wanna say "I told you so", but I told you so. Quote: "Ilya Sutskever, co-founder of AI labs Safe Superintelligence (SSI) and OpenAI, told Reuters recently that results from scaling up pre-training - the phase of training an AI model that uses a vast amount of unlabeled data to understand language patterns and structures - have plateaued." ... threads.net/@yannlecun/pos…
Where's the American equivalent of these? I don't think even Boston Dynamics has this kind of maneuverability. It seems so flawless. It edges on... Making me feel as if... These machines are alive, the way they self-correct. Balance. Plan and prepare their next movements. A… Less
Where's the American equivalent of these? I don't think even Boston Dynamics has this kind of maneuverability. It seems so flawless. It edges on... Making me feel as if... These machines are alive, the way they self-correct. Balance. Plan and prepare their next movements.
Again, I don't see this in any... mainstream American robotics company right now. Maybe we have all of this behind the scenes, but... it's just very weird to see this coming from random Chinese startups. If we have these behind the scenes, what does China have behind the scenes?
probably a SOTA coder if none of the current SOTA labs release something new
then again, it’s about that time for OpenAI’s early Christmas present… 5o?
low hopes for anthropic since watching dario’s interview, especially knowing that we won’t be seeing opus 3.5 anytime soon
if a government doesn't prioritize intelligence and its own hardware, it'll end up beholden to others. it's already happening, and the difference between prosperity and reliance will only increase (probably exponentially if we're looking at intelligence growth as the main factor)
i think we'll start seeing AAA studios buying out indie games just like monopolies buy out startups.
Unfortunately for them though, I think they'll "polish up" the game after takeover, ironically making it worse after going through the AAA studio "upgrade / expand" filters.
i downloaded zen browser, used it for 3 minutes, and closed out of it. arc is still king for my use cases. maybe they will open source it? or their new v2 paid version of it will come out soon? idc if i have to pay $10 or whatever. it's a daily driver, why wouldn't I?
it's insane to me that o1-preview is 5x more expensive for input tokens and 4x more for output tokens compared to claude 3.6 sonnet...
even though sonnet 3.6 is better (at least for coding/creativity).
now anthropic please, then i can remove 80% of custom API code in systemsculpt and just use openai endpoint schema for everything https://t.co/UgGHSe4azM
seriously considering removing all of my AI service provider options from SystemSculpt (currently: OpenAI, Groq, Anthropic, Grok, OpenRouter, Local)... and just making it super simple: OpenRouter and Local. it's becoming more mainstream, has all of the latest models from all of… Less
seriously considering removing all of my AI service provider options from SystemSculpt (currently: OpenAI, Groq, Anthropic, Grok, OpenRouter, Local)...
and just making it super simple: OpenRouter and Local.
it's becoming more mainstream, has all of the latest models from all of the top providers, a bunch of free models, little to no rate limits vs. native model use...
hopefully they fix this repetitiveness that often (very often) occurs specifically with SearchGPT. hey @OpenAI @sama I love it otherwise; it's my daily search driver since day 1 and I'm not looking back, only forward to even better improvements/evolution.
i used to drink nearly 2g of caffeine a day - no, not mg, not 2g of coffee, 2g of actual, pure caffeine - mostly in the form of white monster energy drinks (140mg caffeine each) and 30-40 grams of nescafe gold instant coffee (~40mg of caffeine in each gram)
using o1-preview, there's a "new version of chatgpt", giving me two reasoning replies. i think full o1 is coming super soon, maybe next week (or maybe today, thursday?)
after trying out @aide_dev by @skcd42 for a 1-file codebase for simple edits, trying out many different strategies of communication to see if my prompting may be the bottleneck, i can confidently say that it's nowhere near cursor-level quality nor efficacy. i do have high hopes … Less
after trying out @aide_dev by @skcd42 for a 1-file codebase for simple edits, trying out many different strategies of communication to see if my prompting may be the bottleneck, i can confidently say that it's nowhere near cursor-level quality nor efficacy.
i do have high hopes for it, or rather, its plan -> execute strategy shown in the sidebar.
i will revisit it in 1 month.
i wonder what's going on on threads LMAO
what a beautiful start to the day.
the worst part of the for you page is that i keep seeing the same shit over and over for the entire day, it feels like it's begging me to like or comment or bookmark it, but no - i will not give in to your ragebait politics slop posts.
if i had $10/m for every time a bot with these same stats followed me (always following over 7K, always under 100 followers), i'd be at $100K MRR
I wonder if o1-mini also falls under these updates, or at least o1 comes with reduced prices.
it would be such a 1-up on anthropic with their recent price INCREASE news, imagine if openai DECREASED their SOTA?
The full version of OpenAI o1 is not far away. o1 will be a huge, It will also include Vision, as we saw a few days ago when it was accidentally rolled out. A few more features were announced for o1 during London Dev Day, and I'm also excited about streaming. … Less
The full version of OpenAI o1 is not far away. o1 will be a huge, It will also include Vision, as we saw a few days ago when it was accidentally rolled out. A few more features were announced for o1 during London Dev Day, and I'm also excited about streaming. x.com/legit_rumors/s…
at this point, crumbl's main selling point is dirty bulking
regardless of what country you are from/residing in, you should love your country. if you don't, which country do you love? why aren't you doing everything in your power to relocate to that country?
if you don't love any country, where's your home? your community? family?
anthropic should have just eaten the extra cost instead of 4xing the price. all anyone is talking about is the price increase.
no one is talking about how it’s better than opus 3.
what a missed opportunity for them. shame.
the worst part of all these llm competitors is that everyone likes to do things with their api a little differently. just bow down to openai's standard which has been around for years, stop trying to make your own anthropic, im looking at you. xAI is openai endpoint compatible. … Less
the worst part of all these llm competitors is that everyone likes to do things with their api a little differently. just bow down to openai's standard which has been around for years, stop trying to make your own
anthropic, im looking at you. xAI is openai endpoint compatible. WHY AREN'T YOU OMG what a pain.
is there seriously no litellm equivalent in npm?
it's so funny seeing people write "i'm gonna delete this app, x sucks, elon ruined it" - where are you gonna go? THREADS? instagram? facebook? LMAO
no matter who wins, twitter is gonna be so funny later this week, im so excited
new look, who dis
(new systemsculpt website live)
I mean come on. I feel that not enough people are talking about how amazing ChatGPT Search is.
latest random news, chatgpt search vs. perplexity. which do you prefer? i prefer chatgpt search.
bro wake up new harambe just dropped 🐿️
1,000,000% @finkd is part of tpot, shitposting as a lowkey anon, making sure he keeps up with the vibes as best he can. you think this amazing ai discourse is happening on threads? FACEBOOK? he'd be stupid not to secretly be part of x's ai community (he's not stupid)
after trying out the new version of github copilot, even with o1-preview as the main driver, i can confidently say... it's shit compared to cursor. long live cursor.
i swear, once im at a stable 10k mrr, im making an on-the-fly x feed filter that removes ragebait/engagement bait/politics from my feed. i know musk won't do it, it'd be stupid of him to do so; that's where most money comes from rn it seems
being concise accelerates goal achievement; being precise enhances output quality. with LLMs this is more true than ever before
it really does seem like the full version of o1 has released (by "mistake", imo). so this week, i think, is going to be super interesting.
white is o1 (full?), black is o1-preview unlike o1-preview which gets this wrong, this assumed o1 (full?) gets it correct. it ALMOST falls into the same gender-identity overthinking trap, but it gets it right in the end (you can see it battling with itself in the thinking proces… Less
white is o1 (full?), black is o1-preview
unlike o1-preview which gets this wrong, this assumed o1 (full?) gets it correct. it ALMOST falls into the same gender-identity overthinking trap, but it gets it right in the end (you can see it battling with itself in the thinking process, lmao)
though, it is important to note: o1 tells me "lol this is an old puzzle", so it might just be pulling the final answer's certainty from training data.
the next era is all about hardware. software is practically solved imo
the more you break down programming steps to an llm, the less errors it will make. seems like common sense, but those that weren't programmers before llms don't understand the build out process as well - they look at a module as one step, where in reality it's 20-100 steps.
im starting "no nibble november", as in i am done nibbling at 30+ projects simultaneously - i am focusing on one main category and we'll see how things look by the end of the month
they need to add a chatgpt search filter to the chat history ASAP, the sidebar is just spammed now with my quick searches and it's only been a few hours so far of use from me
shopping is already a seamless experience with chatgpt search btw, don't you think they'll make bank on the conversions from chatgpt -> straight into the website to buy through affiliate/referral/marketing id links?
one chatgpt search downside i've noticed (though it's more annoying than "wrong")
- it repeats itself on followup queries
- it still answers the new question and finds new sources again, so that's good
- seems like an easily solvable problem to de-duplicate
i always wonder how much planning goes into cool little one-off halloween/etc holiday fun videos like this
like, was this always meant to be part of the functionality, or did brett think "this would be cool for halloween, MAKE IT HAPPEN ENGINEERS!"
and will it be useful again?
yeah... im cancelling my perplexity subscription, openai's search is game changing.
Luce is going to be the largest positive long-term influence for catholicism from the vatican since... probably since constantine's conversion. it's going to take catholicism to (even higher) mainstream levels than ever previously imagined.
few.
ok, how about - get this - christianity?
new net worth goal: $1 decillion
insane to me that @cursor_ai STILL doesn't have a commit message generator button like copilot does. is it patented or something?! seems like a no-brainer feature, and also seems like cursor deliberately isn't adding it in for some reason.
probably because python is the best performing language for LLM coding success
