Tofunori

529 posts

Tofunori

Tofunori

@thielaut

Master’s student in Glaciology at University of Trois-Rivières. Passionate about remote sensing, computational modeling, LLM, and climate impacts on glaciers.

Trois-Rivières, Québec Katılım Ekim 2024
403 Takip Edilen47 Takipçiler
Tofunori
Tofunori@thielaut·
@gdb Chatgpt Work slipped into my research workflow with almost no friction: code analysis on my mixed-effects models, sanity-checking DuckDB panels, drafting in both FR and EN. The win isn't "writing faster", it's not losing the thread across a dozen files.
English
1
0
0
111
Greg Brockman
Greg Brockman@gdb·
Was very cool to hear about the reasons people love Sol. We're doing the promotion again, except this time for ChatGPT Work: Tweet what you love about ChatGPT Work, claim $100 in free credits, get more work done. First 10k get the free tokens: share-chatgpt-work.openai.chatgpt.site
Tibo@thsottiaux

Or… what if we gave you $100 in Codex credits if you tell us what you love about GPT-5.6 Sol or why you switched? Tweet it, claim your gift, enjoy more usage. First 10k get the free tokens! switch-to-codex.openai.chatgpt.site

English
5.4K
986
4K
964.1K
Command Code
Command Code@CommandCodeAI·
Who wants access to Qwen 3.8 model (2.4T params)?
English
157
15
781
43.1K
Tibo
Tibo@thsottiaux·
What are you building this weekend?
English
2.6K
63
3.3K
537.3K
Tw93
Tw93@HiTw93·
想和大伙分享一下，最近我用自己产品 Mole 遇到的 4 个 Aha Moment，对我使用非常有帮助。 第一个是，有一天下午居然收到了一个消息通知，说我的 Airpods 电量不足要充电了，这个功能惊喜到了我，是当时做电池健康的时候 AI 主动给我加上的，Mole 会提醒和Mac关联的蓝牙带电池的设备低电量充电，但是我一直没有遇到这个场景，直到遇到了，感觉非常贴心，但频繁不打扰。 第二个是，对于每日使用的我居然还可以清理掉 86G，之前有一个版本加上清理项目里面肯定无用的build编译内容，我的 Rust 项目产生了不少 target 内容，一下子清理这么多，非常有满足感。 第三个是，屏幕常亮的功能，经过用户提醒给加上了3种不同的常亮行为，我还是喜欢把屏幕打开的方式，比如周末你突然出门，你的 AICoding 可以用手机客户端很好的控制让他干活，非常稳定，也帮我节约了不少异步的时间。 第四个是，Mole 的电量显示居然也支持展示你的 iPhone 的电量了，之前这一块不是很好实现，但是很巧妙的实现了，现在可以在状态里面很清楚看到与你相关的设备的电量情况，也不打扰，但是很安心。 更神奇的是，以上这4个功能的起源都不是我想到的，是初期使用 Mole 的朋友反馈告诉我要加的功能，我加上后对我帮助非常大，非常感谢！话说你使用 Mole Mac 有惊喜的地方是哪一些，以及你非常建议 Mole 要加的功能，欢迎告诉我。 mole.fit
Tw93 tweet mediaTw93 tweet mediaTw93 tweet mediaTw93 tweet media
中文
79
9
282
42K
Tofunori
Tofunori@thielaut·
@iruletheworldmo I'm exhausted! Fable 5 really unblock something and now every lab is trying to close the gap. What a time!
English
0
0
2
874
🍓🍓🍓
🍓🍓🍓@iruletheworldmo·
‼️ two bits of breaking news for loyal strawberries. 1. openai’s next two models have finished training and are named nova and quasar. this will be the public name of the models. 2. ilya and ssi have reached a stage they need to go public with their models, they will release a frontier level model in august this year if you thought the last six months were fast.
English
96
67
1.4K
102K
Tofunori
Tofunori@thielaut·
@theo Yep it's been a while no?
English
0
0
0
73
Theo - t3.gg
Theo - t3.gg@theo·
Any resets coming soon? 💀
Theo - t3.gg tweet media
English
128
19
1.9K
136.4K
Tofunori
Tofunori@thielaut·
@chetaslua It's good! But the one from Opus 5 has more nuances in the clouds etc
English
1
0
0
1.3K
Chetaslua
Chetaslua@chetaslua·
Holy Shit Kimi 🤯 the more i test it the more i am thinking , opensource and china is not that behind from current gen SOTAs this is better output than upcoming opus 5 same prompts and different result
Chetaslua@chetaslua

🚨Exclusive Opus 5 first output was seen as claude honeycomb EAP in cursor and removed after two hours so we can have opus-5 AS EALY AS week or next week while it was there i ran two prompts and i also tried to test my jailbreak on it > this leads to fallback to opus 4.8 , and we know claude started to add this switch features in all latest models where harmful stuff get routed to weak model this means Honeycomb is bigger and better model than 4.8 and worse then fable 5 as i have two test ran already here are the tests that i ran on them

English
46
55
978
226.4K
Tofunori
Tofunori@thielaut·
@thsottiaux A banked reset will be a good start of the day! Having so much fun with Sol
English
0
0
2
186
Tibo
Tibo@thsottiaux·
Embarrassment of riches. But looks like we might hit 9M soon. Should we reset the ChatGPT Work and Codex usage again or give it some space?
English
4.7K
481
15.3K
3.6M
Tofunori
Tofunori@thielaut·
@thsottiaux @LexnLin I love Sol because it’s a model I can trust to dig deep into my code for my thesis and catch bugs other models miss.
English
0
0
0
10
Tibo
Tibo@thsottiaux·
Here you are! Thinking I am about to announce a reset. But no. I’m just scrolling twitter and looking for feedback on ChatGPT Work. What should we improve?
English
3.3K
156
8.6K
817.6K
Tibo
Tibo@thsottiaux·
Tomorrow might be 8M active user celebration day. Just saying
English
1.6K
402
12.3K
2.9M
Tofunori
Tofunori@thielaut·
@ajambrosino word wrap for LaTeX files, and the ability to annotate files opened from the Files tab, please 🙏
English
0
0
0
242
Andrew Ambrosino
Andrew Ambrosino@ajambrosino·
team is cooking some great desktop improvements
English
97
16
1.1K
41.3K
Tofunori
Tofunori@thielaut·
@sama Wow Anthropic needs to learn from this!
English
0
0
0
7
Tofunori
Tofunori@thielaut·
@rezoundous that's for sure.. and will come back when Fable 5.1 is out
English
0
0
0
139
Tyler
Tyler@rezoundous·
Fable might be removed after Opus 5 gets released.
English
25
2
126
9.1K
Tibo
Tibo@thsottiaux·
Updates for Codex and ChatGPT Work users. No nerfing, only good stuff! - We have landed inference optimizations and are passing down savings to all the subscriptions for GPT-5.6 Sol. That should result in around 10% more usage on its own. - We noticed that by changing the context size limit in the product to 372k for GPT-5.6 Sol, up from 272k for GPT-5.5, it resulted in more usage being charged than intended. We have reverted to 272k and will work to roll back out to 372k in the days to come. You should notice that usage drains significantly less after this change. - To understand where the extra usage was coming from, we ran some experiments where reasoning efforts were changed (referred to as juice values under the hood) and have reverted this. - There is slightly more usage of multi-agent than intended in high and xhigh reasoning effort, we are fixing this going forward. Also fixing a small other thing we noticed with auto-review where we can be more efficient. And we continue to have the 5h limit temporarily not apply. Enjoy the rest of the weekend!
Fix@FixlationAI

OpenAI has reduced GPT-5.6 Sol's thinking budgets in an effort to make the model more efficient They essentially bumped everyone's reasoning down by 1... so if you were running Sol Extra High, you now have to set it to Max to get the same effort So we basically don't have Max reasoning anymore, how do you feel about these changes? 🤔

English
1.1K
610
9.9K
2.2M
Barathwaj Anandan
Barathwaj Anandan@BarathAnandan7·
@thielaut @thsottiaux They undersell Luna for some reason. Luna is a work horse. 5x cheaper and can do 95% for me tbh. I move to sol solely when I know i need it :) they built beautiful models. Luna is same or little better than glm 5.2 for my work!
English
2
0
3
200
Barathwaj Anandan
Barathwaj Anandan@BarathAnandan7·
I barely sleep anymore because of Codex… and I’m not even mad about it. 😭 I just wanted to say Codex and @thsottiaux have genuinely changed my life. 5.6 ( mainly you, Luna!) is such a game changer. Not just because it’s insanely capable, but because of the empathy behind it, the way @OpenAIDevs listen, fix issues quickly, and actually let people cook. I look forward to my Codex resets like I used to wait for Amazon packages to arrive. Btw the 5 hour limit is gone.. enough touching grass, go COOK!!
English
13
7
265
15.1K

Keşfet

@gdb @CommandCodeAI @thsottiaux @HiTw93 @iruletheworldmo @LexnLin @iamsaranhere @theo