muzz khan (@muzzdotdev) header
muzz khan (@muzzdotdev) profile picture

(l,y)earning

Joined July 2023
Photo by muzz khan (@muzzdotdev)Photo by muzz khan (@muzzdotdev)Photo by muzz khan (@muzzdotdev)Photo by muzz khan (@muzzdotdev)Photo by muzz khan (@muzzdotdev)Photo by muzz khan (@muzzdotdev)Photo by muzz khan (@muzzdotdev)Photo by muzz khan (@muzzdotdev)Photo by muzz khan (@muzzdotdev)Photo by muzz khan (@muzzdotdev)
muzz khan retweeted
Oh my god there’s a bug! And this one’s bad! I can’t believe it! Software has vulnerabilities! This is so crazy! Ahhhhhhh! Why are yall like this.
29
14
1
199
6,023
muzz khan retweeted
i sent this investor update super late at night and accidentally attached the wrong chart. it only goes till september. sorry for the inconvenience caused. here’s the right one:
AI startups grow fast.
83
28
19
880
143,976
muzz khan retweeted
99
2,648
79
37,146
455,152
muzz khan retweeted
i love her so much i could cry
53
1,425
36
29,365
150,092
uma hora você é o padeiro na outra você é o pão
57
8,745
84
80,264
609,822
muzz khan retweeted
Replying to @OpenAI
yo guys, you have a bug in your metric intervals go ahead and add a rollover
1
2
169
muzz khan retweeted
Greptile has reviewed over 395,000 pull requests across @NVIDIA‘s large, complex codebases, where keeping up with code review was taking a growing share of engineers' time. On one team, the average time to merge went from more than 24 hours to 6 hours.
4
8
4
111
39,428
she sits weird AND meows ALL DAY
64
2,373
139
29,279
1,023,468
muzz khan retweeted
greptile's mission is to automate code validation. i want share more about what that means, and why we're working on this problem.
11
5
4
160
30,266
1
92
4
932
17,809
muzz khan retweeted
introducing greptile's free tier. we've made a lot of improvements over the past few months to greptile, and we want to share those with more developers. you can now get 50 reviews per month as a solo developer for free
67
31
12
753
90,321
when i call my cat fat, i am not fatshaming him. in fact, i am fatpraising him.
58
7,125
205
66,555
690,012
muzz khan retweeted
How I feel every time I get a fever
26
4,936
87
36,101
427,777
34
4,694
122
32,297
645,498
Excited to see the use of GEPA-optimized LLM judges for data filtering in MAI-Thinking-1 model's pre-training pipeline!
Super excited to announce seven new world-class MAI models today. They represent what we consider a new era in AI designed to keep you in control and on the frontier. First is our text foundation model, MAI-Thinking-1, exceptionally strong on reasoning and SWE tasks. - It’s a 35B active parameter MoE with a 256K context window. Independent human raters on Surge prefer it for overall quality in blind side-by-sides versus Sonnet 4.6, and it’s achieved 97% on AIME 2025, the key measure of its general-purpose reasoning abilities. - It's at 53% on SWE Bench Pro, placing it right alongside Opus 4.6 on one of the toughest coding benchmarks. - And since we co-designed our models with our own silicon, MAI-Thinking-1 is optimized on our MAIA 200 chip. Benchmarking head-to-head against the GB200, we see 30% better performance per dollar as well as a 1.4x performance-per-watt gain when running our MAI models on the MAIA 200 end-to-end. Next is MAI-Image-2.5 and its Flash variant. Two super strong models now at #2 on the leaderboards, surpassing the score of Nano Banana 2 on image editing. Last for now is MAI-Code-1-Flash, our new inference efficient coding model, especially tuned for VS Code and GitHub Copilot CLI. - Code-1-Flash achieves 51% on SWE Bench Pro, despite having just 5B parameters, putting it closer to Haiku in size but cheaper in cost. All of this is the foundation for Microsoft Frontier Tuning. It lets you customize our models to create custom, company-specific agents that only you control. You can make our model, your model. Your data. Your agents. Your moat. Early adopters are already seeing a difference. When we tuned our models for McKinsey’s tasks, MAI delivered the highest win rate, outperforming GPT-5.5 on quality, while being 10x lower on cost. Also really excited to be collaborating with the amazing team at Mayo Clinic to jointly train a new frontier AI model for healthcare. Our announcements today mark another milestone on the road to humanist superintelligence. You can learn more and about our other new models in our latest blog: microsoft.ai/news/building-a…
3
22
5
168
52,876
Greptile is coming to Tokyo, Japan! 🇯🇵🦎 To celebrate, we are hosting a community Happy Hour for our users and friends. It's happening on Monday, June 8th, 7-10 pm. Get a chance to win free Greptile credits, merch, and free drinks!! Looking forward to it. :) ----------------------------------------------------------- @greptile が東京にやってきます!🗼 これを記念して、ユーザーの皆さま・お友達のためのコミュニティ・ハッピーアワーを開催します。 📅 6月8日(月) 🕖 19:00–22:00 Greptileクレジットやオリジナルグッズが当たるチャンス、さらにドリンクも無料です! 皆さまにお会いできるのを楽しみにしています😊 RSVP/参加登録: luma.com/o8z4d4nj
3
10
4
63
982,853
muzz khan retweeted
Sorry, I prematurely abstracted. This doesn’t usually happen.
23
24
4
773
40,724
muzz khan retweeted
launching 5 things: 1. multi-repo context support 2. rebuilt web app for super large orgs 3. integrations with claude/codex/devin 4. .greptile/rules files 5. rebuilt learning so greptile maintains internal docs about your company
18
10
6
126
24,158
muzz khan retweeted
more than a quarter of all code reviewed by greptile is now written by "background agents" - completely autonomous e2e coding agents like devin. we analyzed millions of PRs written by background agents and compared their - code churn rates - revert rates - bugs per PR against human baseline. read the full post on our blog!
4
18
56
5,014
muzz khan retweeted
Me asking claude if devin fixed the bug greptile found
1
1
16
692