Mohamed Zayed

Research Engineer @ my journey of 10K hrs,
Opinions are my own.

[Mohamed Zayed @MoZayed007]

@mervenoyann @andimarafioti @micuelll @orr_zohar Congrats 👀

[Mohamed Zayed @MoZayed007]

@jietang the ability to handle a swarm of delegated subagents, basically kimi swarm is so good yet so expensive. making this efficient , scalable and standard will help a lot 🙏💙

[Mohamed Zayed @MoZayed007]

@gpusteve Happy birthday, hope you enjoy many years of successful wafer 🎂

[Mohamed Zayed @MoZayed007]

I hope it's not benchmaxing will try to use it on my personal projects.

output tokens quality should have its own eval system, there should be a way to systematically define the frontier taste of a model other than domain specific benches hill-climbing.

[Datacurve @datacurve]

GLM 5.2 is now on DeepSWE as the top open-source model on our leaderboard.

With a pass@1 score of 44% at max effort, GLM 5.2 is indisputable #1 open-source model besting Kimi K2.7 Code by 17%.

[Mohamed Zayed @MoZayed007]

@poteto please a poteto wearing a space suit

[Mohamed Zayed @MoZayed007]

@YouJiacheng Thanks for giving me something to derive and study during the weekend 🫡

[Mohamed Zayed @MoZayed007]

@_xjdr I knew I'm following and learning from the right person, and I was right, kimi is a lovely model 🩵🫡

[Mohamed Zayed @MoZayed007]

@xeophon Moonshot is the lab I'm feeling butterflies from when studying their released projects, knowledge, papers, etc.

[Mohamed Zayed @MoZayed007]

waiting for k3 being animated in a video with a phonk background hill-climbing over all models 💙🏎️ or Max Verstappen song either works 👌

[Cognition @cognition]

Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers.

Models write sloppy code that works but isn’t maintainable. Our eval is first to measure: would you actually merge this code?

[Mohamed Zayed @MoZayed007]

@crystalsssup @KimiDevs I belive the next release will at least 10x growth rate.
waiting with positive and happy thoughts for kimi 💙🏎️

[Mohamed Zayed @MoZayed007]

I'll keep hoping from my deepest wishes 🤞

[MATS Research @MATSprogram]

🚨 Applications for MATS Autumn 2026 close tonight (June 7 AoE)!

Spend 10 weeks fully funded working with mentors from Anthropic, DeepMind, OpenAI, Redwood Research, SecureBio, and more.

New this cohort:
🧬 Biosecurity
🚀 Founding & Field-Building

Apply now: https://t.co/XWIdre04C8

[Mohamed Zayed @MoZayed007]

I got the urge to convert all my util, tooling projects, routine stuff I built for self satisfaction take all that , add an agentic interface / module suite like mcps, profiles/personas as yaml if needed and connect them to Hermes , opencode , kilo cli and it feels so good 💙👌

[Mohamed Zayed @MoZayed007]

@KimiDevs Hello love <3

[Mohamed Zayed @MoZayed007]

@thdxr If the decision-maker always hurdles and blocks 2 of the 3 when we are still at poc level (never held a knife managing a kitchen kind of situation), shall I switch companies?

[Mohamed Zayed @MoZayed007]

@poteto Thanks for the great potato 🥔

[Mohamed Zayed @MoZayed007]

@real_deep_ml https://t.co/z3f3jo57KF
Using whatever is possible to support, for example, Triton

[Mohamed Zayed @MoZayed007]

@real_deep_ml waiting for other podcasts 🙏

[Mohamed Zayed @MoZayed007]

@lqiao @FireworksAI_HQ done and praying 🫡🙏☕

[Mohamed Zayed @MoZayed007]

@lqiao @FireworksAI_HQ so what about a +1 MTS fit who did communications and Information Engineering (AI concentration), worked on cybersecurity gigs for 3yrs before graduation , and currently working as AI RnD eng. at an ed-tech startup, super passionate towards fireworks tbh