@ZeMariaMacedo speaking of independent thinking - guess I could’ve done better last week highlighting kimi (Moonshot AI) team’s original research 😆
English
Dana
490 posts
Dana
@DanaBuidl
exploring and scaling the frontier. committed to being the change I wish to see. seeking pmf & deep self-awareness. views are my own. NFA.
Katılım Ekim 2018
1.3K Takip Edilen9.4K Takipçiler
Dana retweetledi
Layoffs Impact -
Lots of chatter around reorgs & layoffs recently. I wanted to build a simple framework to understand the financial impact.
All scenarios below are purely hypothetical. Key Observations:
1/ SaaS is far more sensitive than Big Tech
Using $CRM as a reference (and likely representative of other SaaS names):
Assume ~$200k average salary per employee
Layoffs can drive:
- ~40% uplift to GAAP EPS
- ~20% uplift to FCF
This is meaningful. Labor is a much larger % of the cost base, so cuts flow through cleanly.
2/ Big Tech: smaller EPS impact, larger FCF leverage
For $META, $MSFT, $GOOGL:
A 20% layoff implies:
- ~5–15% uplift to EPS
- Significantly larger impact on FCF
The key difference: these companies are already highly profitable, so incremental savings show up more clearly in cash than in earnings.
Take $GOOGL as an example: At $200k–$800k salary per employee:
- FCF could increase by ~30–80%
- Equivalent to enabling ~10% incremental CapEx deployment
In other words, layoffs don’t just boost margins - they can directly fund the AI capex cycle.
---
To ground the analysis, here are the base assumptions:
- 20% workforce reduction
- 15% tax rate
- $200k–$800k salary range per employee
Full analysis: open.substack.com/pub/robonomics…
English
now it’s mostly engs. we need next level intelligence to empower literally everyone to become full stack.
and I look forward to the day (hopefully soon) when we’re reduced to taste, creativity, and judgment.
The All-In Podcast@theallinpod
Jensen Huang on the future of coding: “Every engineer is going to have 100 agents.” Jensen: “Everything that's too big, too heavy, takes too long, those ideas are all gone.” Chamath: “You're reduced to creativity. Like, what can you come up with?” Jensen: “Exactly.” “Now the question is, how do you work with these agents?” “Well, it's just a new way of doing computer programming.” “In the past, we code.” “In the future, we're going to write ideas, architectures, specifications. We're going to organize teams. We're going to help them define how to evaluate the definition of good versus bad. What does it look like when something is a great outcome? How to iterate with you, how to brainstorm.” “That's really what you're looking for.”
English
Dana retweetledi
We've evaluated a lot of base models on perplexity-based evals and Kimi k2.5 proved to be the strongest!
After that, we do continued pre-training and high-compute RL (a 4x scale-up).
The combination of the strong base, CPT and RL, and Fireworks' inference and RL samplers make Composer-2 frontier level.
It was a miss to not mention the Kimi base in our blog from the start. We'll fix that for the next model.
Kimi.ai@Kimi_Moonshot
Congrats to the @cursor_ai team on the launch of Composer 2! We are proud to see Kimi-k2.5 provide the foundation. Seeing our model integrated effectively through Cursor's continued pretraining & high-compute RL training is the open model ecosystem we love to support. Note: Cursor accesses Kimi-k2.5 via @FireworksAI_HQ ' hosted RL and inference platform as part of an authorized commercial partnership.
English
Dana retweetledi
Congrats to the @cursor_ai team on the launch of Composer 2!
We are proud to see Kimi-k2.5 provide the foundation. Seeing our model integrated effectively through Cursor's continued pretraining & high-compute RL training is the open model ecosystem we love to support.
Note: Cursor accesses Kimi-k2.5 via @FireworksAI_HQ ' hosted RL and inference platform as part of an authorized commercial partnership.
English
thanks for the full list!
@Kimi_Moonshot mentioned👀
Deedy@deedydas
Every single one of the 103 companies Jensen called AI Native today.
English
Dana retweetledi
Elon also thinks @Kimi_Moonshot’s work is impressive 👀
Elon Musk@elonmusk
@_avichawla Impressive work from Kimi
English
Dana retweetledi
Jianlin Su and kexue.fm are all you need for architecture innovation.
Yu Zhang 🐙🌘@yzhang_cs
The idea of rotating attention by 90° is sooooooo cool (credits to @Jianlin_S 's insights), and it surprisingly works. We (w/ the amazing @nathan) are so excited about this— been working on the paper for months and couldn't stop. Go give it a try. It's a drop-in replacement for standard residuals, born in 2015. really like the figs btw :-)
English
Dana retweetledi
unbelievably impressive 🤩
Kimi.ai@Kimi_Moonshot
Introducing 𝑨𝒕𝒕𝒆𝒏𝒕𝒊𝒐𝒏 𝑹𝒆𝒔𝒊𝒅𝒖𝒂𝒍𝒔: Rethinking depth-wise aggregation. Residual connections have long relied on fixed, uniform accumulation. Inspired by the duality of time and depth, we introduce Attention Residuals, replacing standard depth-wise recurrence with learned, input-dependent attention over preceding layers. 🔹 Enables networks to selectively retrieve past representations, naturally mitigating dilution and hidden-state growth. 🔹 Introduces Block AttnRes, partitioning layers into compressed blocks to make cross-layer attention practical at scale. 🔹 Serves as an efficient drop-in replacement, demonstrating a 1.25x compute advantage with negligible (<2%) inference latency overhead. 🔹 Validated on the Kimi Linear architecture (48B total, 3B activated parameters), delivering consistent downstream performance gains. 🔗Full report: github.com/MoonshotAI/Att…
English
Dana retweetledi
The idea of rotating attention by 90° is sooooooo cool (credits to @Jianlin_S 's insights), and it surprisingly works.
We (w/ the amazing @nathan) are so excited about this— been working on the paper for months and couldn't stop.
Go give it a try. It's a drop-in replacement for standard residuals, born in 2015.
really like the figs btw :-)
Kimi.ai@Kimi_Moonshot
Introducing 𝑨𝒕𝒕𝒆𝒏𝒕𝒊𝒐𝒏 𝑹𝒆𝒔𝒊𝒅𝒖𝒂𝒍𝒔: Rethinking depth-wise aggregation. Residual connections have long relied on fixed, uniform accumulation. Inspired by the duality of time and depth, we introduce Attention Residuals, replacing standard depth-wise recurrence with learned, input-dependent attention over preceding layers. 🔹 Enables networks to selectively retrieve past representations, naturally mitigating dilution and hidden-state growth. 🔹 Introduces Block AttnRes, partitioning layers into compressed blocks to make cross-layer attention practical at scale. 🔹 Serves as an efficient drop-in replacement, demonstrating a 1.25x compute advantage with negligible (<2%) inference latency overhead. 🔹 Validated on the Kimi Linear architecture (48B total, 3B activated parameters), delivering consistent downstream performance gains. 🔗Full report: github.com/MoonshotAI/Att…
English
@DanaBuidl @yzilabs @BinanceLabs Wishing all the best to you, Dana!
Thanks for the trust and for being such an inspiration to all of us. Can’t wait to see your next chapter 🫶
English
yesterday was my last day with @yzilabs (formerly @BinanceLabs). reflecting on nearly 4 years, I am deeply grateful for the journey. looking back, there are four highlights that I will carry with me:
1/ it has been a privilege to work closely with and learn from our LPs @cz_binance @heyibinance. thank you for the trust, the mentorship and the opportunity.
2/ to the founders who let me into your journeys (in chronological order):
- in 2023, @tn_pendle (@pendle_fi), @gdog97_ (@ethena)
- in 2024, @0xpengyu (@ParticleNtwrk), @no89thkey (@brevis_zk), @annakaz & @artieart88 (@vana), @realyanxin (@Sign), @HanyangWang (@Funes_World)
- in 2025, @ohyishi (@OneKeyHQ), @0xZergs (@USDai_Official), @BlockRazor_Inc, @thecryptoskanda & @heroyangv (@Hertzflow_xyz), the "leos" in crypto @Leozayaat (@42space) @leokeeplearning (@apro), and others I’ve had the honor to support.
the list is far longer than one post allows, but I’m wishing only the best for all of you.
3/ driving the acquisition of Jackson @0xOar and his team into @BinanceWallet with @Winson_Bit's support. seeing that deal through was a testament to the persistent hard work and the relentless pursuit of excellence in this space.
4/ last but certainly not least, I’m grateful to have worked with super talented, serious and hardworking investors: @cyodyssey, @p2fp2e, @mc_9393, @The_Babylonians, and @odagius. over the years, many efforts and hearts have been poured into the work. some ventures flourished and some didn’t, yet we either gained or learned. 相信功不唐捐。
I’m equally proud of the talent I brought into the team, including @0xRickyW, @yunna025, and @Hui_Huangg (referral credit to @gizmothegizzer - another wonderful female founder we backed in 2024). with the best intern @cz_binance and with @ellazhang516 leading the team with vision, @yzilabs will only get better.
my conviction in this space remains unchanged:
- btc remains a store of value. 十币称侯 x.com/ohyishi/status…
- crypto as an asset class is here to stay, rooted in its unique risk/return profile and permissionless nature.
- blockchain as an open, borderless infrastructure will harness more efficient financial rails and provide universal access to on-chain assets and high-throughput, low-cost services.
I’m moving on to new endeavors, but I’ll continue to cheer for everyone from the sidelines.
blockchain doesn’t need permission to be useful and to thrive.
long live crypto.
青山不改，绿水长流，我们后会有期。
ps:
to all who celebrate - 元宵节快乐，马年越来越好！
English
@odagius @yzilabs @BinanceLabs hahaha we’ve come a long way 🤣🙌 thank you for always being such a steady, understanding, and supportive friend. i know that you, my fellow ex-gs/labs friend, are going to crush it too. wishing you nothing but the absolute best!
English
@DanaBuidl @yzilabs @BinanceLabs Loved working with you from the very start (even those crazy days when the whole team was only 3 people haha). Can’t wait to see what comes next! Wherever you go next, you’re going to crush it 🚀
English