Sidharth Sirdeshmukh

192 posts

Sidharth Sirdeshmukh banner
Sidharth Sirdeshmukh

Sidharth Sirdeshmukh

@sidharf

i'll take my salary in tokens please

Boston, MA Katılım Ağustos 2023
29 Takip Edilen17 Takipçiler
Sabitlenmiş Tweet
Sidharth Sirdeshmukh
Sidharth Sirdeshmukh@sidharf·
@cursor_ai Token pricing varies wildly across the models, so I plotted my long-horizon evals against $ cost. Frontier shifts big time
Sidharth Sirdeshmukh tweet media
English
2
0
37
8.5K
Lee Robinson
Lee Robinson@leerob·
Yep, Composer 2 started from an open-source base! We will do full pretraining in the future. Only ~1/4 of the compute spent on the final model came from the base, the rest is from our training. This is why evals are very different. And yes, we are following the license through our inference partner terms.
Fynn@fynnso

was messing with the OpenAI base URL in Cursor and caught this accounts/anysphere/models/kimi-k2p5-rl-0317-s515-fast so composer 2 is just Kimi K2.5 with RL at least rename the model ID

English
308
148
2.3K
928.8K
Sidharth Sirdeshmukh
@thdxr I think it's actually 5.3 Codex under the hood. Been running some experiments past 2 weeks and can share evidence
English
0
0
0
357
Sidharth Sirdeshmukh
Some more evidence supporting my hypothesis from the past few hours. These Kimi 2.5 fish tanks (specifically the fish, pebbles, and labeling/ UI) I generated look nothing like the Codex 5.3 and Composer 2 fish tanks in my original article...
Sidharth Sirdeshmukh tweet mediaSidharth Sirdeshmukh tweet mediaSidharth Sirdeshmukh tweet mediaSidharth Sirdeshmukh tweet media
English
0
0
0
53
Yuchen Jin
Yuchen Jin@Yuchenj_UW·
Cursor’s Composer 2 is likely built on Kimi K2.5. The model URL + tokenizer are strong signals. I love this direction: companies mid-train and post-train on top of OSS LLMs. Prediction: open-source model labs will monetize by taking a cut when others build on top of their models and scale to millions of real users. They will enforce this via licensing. That’s the flywheel. That’s how open-source AI thrives.
English
42
19
518
65.6K
Sidharth Sirdeshmukh
@Yuchenj_UW I see your post is speculative too. Couldn't the original account have photoshopped/ doctored the post?
English
2
0
0
232
Sidharth Sirdeshmukh
@TobiasVinc57321 They may have gotten the weights through a partnership w/ Open AI. And, the model got ‘worse’ because they RL’d for efficiency (cost/ time) optimization in coding tasks specifically (GPT base models are good at a lot of other things)
English
1
0
0
75
Tobias
Tobias@TobiasVinc57321·
@sidharf Distill is not the same as a RL'd version How would cursor manage to get gpt 5 weights? How is the resultant model worse? Lol
English
1
0
0
98
Alexander Doria
Alexander Doria@Dorialexander·
If true, they likely just paid Moonshot for a custom license (everything is negotiable). But still an added value to own the full pretrain.
English
3
1
19
1.7K
Mark Kretschmann
Mark Kretschmann@mark_k·
Cursor Composer 2 appears to be built on Kimi K2.5 as the base model! 🤫 The Kimi model was then post-trained further with reinforcement learning for coding performance. I think it's quite likely true, as Cursor wouldn't train a completely new foundation model.
Mark Kretschmann tweet media
Fynn@fynnso

was messing with the OpenAI base URL in Cursor and caught this accounts/anysphere/models/kimi-k2p5-rl-0317-s515-fast so composer 2 is just Kimi K2.5 with RL at least rename the model ID

English
45
42
801
100K
Ethan Mollick
Ethan Mollick@emollick·
My experience so far with LLM fiction writing is that it takes advantage of our assumption that an author is writing things for a reason, so we are charitable to a book's quirks & do mental work to assign them real meaning. But the AI doesn't have a reason, its just bad writing.
English
36
14
203
20.1K