Max

3.6K posts

Max banner
Max

Max

@maxxrubin_

Aerospace on pause, AI on purpose 😌

Katılım Mayıs 2022
3.2K Takip Edilen1.1K Takipçiler
Tibo
Tibo@thsottiaux·
@nicdunz holy codex you mean
English
13
2
223
4.5K
Max
Max@maxxrubin_·
@thsottiaux @gabriel1 “Codex, can you make an automation to remind me in three months to mark those words?”
Max tweet media
English
0
0
0
81
Max
Max@maxxrubin_·
Codex team is shipping in every sense of the word Thank you @OpenAI @OpenAIDevs for the super bowl easter egg merch drop 🫡 You can just build things.
Max tweet mediaMax tweet mediaMax tweet mediaMax tweet media
English
2
0
25
637
adi
adi@adonis_singh·
@maxxrubin_ i dont watch superbowl and didnt see their ads, or if i did i barely remember them lol
English
1
0
7
630
Max
Max@maxxrubin_·
@BrianRoemmele It’s the same reason automakers design the front ends of cars to resemble faces! They use the fascia to characterize the car as happy/aggressive/friendly/etc, and that lets us emotionally connect with it by giving the car a more legible ‘temperament’
English
0
0
0
93
Brian Roemmele
Brian Roemmele@BrianRoemmele·
“Why do they need to have a head like humans, it’s pointless. We don’t need it to look like us…” Yeah sure…
English
539
288
2.6K
378.1K
Max
Max@maxxrubin_·
@Dimillian I know it’s kinda out there, but maybe something gdrive-esque for organizing projects/threads (folders/subfolders)? It’s much easier to keep track of ideas/threads spatially as opposed to lists of chats It could just be an optional thing for power users!
English
1
0
1
106
Thomas Ricouard
Thomas Ricouard@Dimillian·
I’m diligently working on enhancing all the features of the Codex app to make it even better for you. If you have any suggestions or ideas for improvement, please feel free to share them with us here!
English
432
11
637
50.5K
Max
Max@maxxrubin_·
@devNEJ Damn that’s forked up
English
1
1
1
399
iamNej
iamNej@devNEJ·
Daddy keh😅😅😂
Indonesia
2.2K
3.7K
134.8K
3.6M
Max
Max@maxxrubin_·
@sama Very thankful that you and your family are safe. Nobody should ever have to go through something like this :(
English
0
0
4
675
Max retweetledi
LM Studio
LM Studio@lmstudio·
Locally AI is joining LM Studio! We are beyond excited to welcome @adrgrondin and @LocallyAIApp to the LM family. Together we are doubling down on native AI experiences across your devices, anywhere you go. Read our announcement lmstudio.ai/blog/locally-a…
English
61
120
1K
124.5K
Max
Max@maxxrubin_·
@nashndiaz new algo went right for the jugular with this one 😭
English
1
0
5
901
kels 🚒
kels 🚒@nashndiaz·
gay friend: I'm gay. depressed friend: I'm depressed. that one gay depressed friend you try to ignore: have you been watching the artemis 2 moon mission updates
English
86
4.2K
33.9K
379.2K
Max
Max@maxxrubin_·
@test_tm7873 4.5 haiku is already criminally underrated tbh
English
1
0
2
34
Locally AI - Local AI Chat
Locally AI - Local AI Chat@LocallyAIApp·
Today, we’re excited to announce that we’re joining @lmstudio! We share a common vision of building amazing products for everyone and making local models more accessible. We’re combining our efforts to bring you the best experience possible.
Locally AI - Local AI Chat tweet media
English
105
129
1.8K
99.5K
Max
Max@maxxrubin_·
@jxnlco but he *is* the chief fun officer??
English
0
0
0
381
Tibo
Tibo@thsottiaux·
I realize yesterday’s Codex reset came in a bit at an unfortunate time given the last one was almost perfectly a week ago. To really celebrate the 3M I’ll reset again tomorrow. Thanks for the feedback!
English
644
298
6.6K
550.4K
Max
Max@maxxrubin_·
@theo It's more tier-comparable to GPT-5.4 Pro imo and it's cheaper (GPT-5.4 Pro $30/$180 vs Mythos $25/$125)
English
0
0
0
1.1K
Theo - t3.gg
Theo - t3.gg@theo·
Claude Mythos is the start of the end. I think this is my psychosis moment.
English
105
911
12.3K
3.3M
Max
Max@maxxrubin_·
@Sauers_ Was working on this a few months ago but only got as far as building some tooling for llama.cpp
English
0
0
1
12
Max
Max@maxxrubin_·
@Sauers_ Instead of training the model to predict its own entropy, give it a tool that returns entropy for the next-token distribution and use GRPO to learn when to call it on verifiable questions and have some penalty for over/under-use
English
1
0
1
37
Sauers
Sauers@Sauers_·
Question: Can I RL Qwen-30B-A3B into correctly predicting how "flat" its own probability distribution will be for the next token of some input? (Sort of like uncertainty / normalized entropy) Answer: No. I tried it w/ GRPO for hundreds of steps and it didn't improve at all 😭
Sauers tweet media
English
7
0
47
2.8K