Biggest ways working in big tech (instagram) has changed in practice in the last 12 months:
- Nearly no handwritten code. i haven't written a single line by hand since March and have shipped features to 100 million+ users.
- Human code review has become rare to non-existent. mainly because it's impossible to keep up. i predict in 2 years we'll look at python the way we look at assembly today.
- Little time on roadmap prioritization. why choose the *best 5* of the 10 projects when you can test all 10 and 3 more you didn't think of before.
- Meeting time has significantly reduced. the opportunity cost of useless meetings is way higher than before.
curious if this matches what others are seeing. will share next what hasn't changed or got worse.
I agree with Dario that we need to pace the frontier. This has been a primary topic of discussions we've had at OpenAI in recent weeks.
Committing to having independent evaluators with employee-like access is a great idea, and we will do the same. We'll have more to share soon.
Japanese startup Preferred Networks is seeking to go public to mass-produce its chips, a reflection of the rising cost and scale needed to stay relevant in a global AI race https://t.co/fuC4dAOIcQ
We're hiring a Research Engineer at @Arsenal ⚽🔴⚪ to work directly with our Men's First Team!
We're building state-of-the-art AI models for the football domain. This role will focus on building the application layer for our research to advance coaching and analysis workflows.
Big news: Kimi-K3 by @Kimi_Moonshot is now #1 in the Frontend Code Arena with 1679 pts, surpassing Claude Fable 5.
This is a 17-place jump from Kimi-k2.6 (#18 -> #1).
In Frontend, Kimi-K3 ranked #1 in 6 of 7 domains: Brand & Marketing, Reference-Based Design, Data & Analytics, Consumer Product, Simulations, and Content Creation Tools, landing #2 only in Gaming behind Fable 5.
The full model weights will be released by July 27.
Congrats to the @Kimi_Moonshot team on this major milestone!
これはいろんなところで言われている話だけど、エー���ェントによる成果物に対する「理解負債」が今新しいボトルネックになりつつある。エージェントは検証能力も高まりつつあり、開発を仮に結構任せたとしてもミスもだんだんしにくくなりつつある中で、人はどうやって開発に参加していくか?というと、それは起こったことを理解することにあると🤔
で、じゃあどうやって理解負債を防ぐか?の具体的なプラクティスが紹介されている:
1. 理解を促すわかりやすいHTMLなりNotionを用意する
2. 直感的に内部を細かく理解できるツールをコーディングエージェントに作ってもらう(記事内だとPrologインタプリタ用のデバッガやゲームを作ってもらっていた)
3. エージェントにNotion上にプランを立ててもらってみんなで議論する
など。
Understanding is the new bottleneck
https://t.co/X55bxTAQW7
Introducing DiffusionBlocks: Block-wise Neural Network Training via Diffusion Interpretation
https://t.co/c9AvsRKybj
What if we didn’t have to hold an entire neural network in memory to train it?
Standard neural net training optimizes all parameters jointly. As a result, the memory required during training grows linearly with the depth of the network.
In our #ICLR2026 paper, we propose DiffusionBlocks, a principled framework to train networks one block at a time, drastically reducing memory requirements while matching end-to-end performance.
With DiffusionBlocks, we split the network into blocks and train them one at a time, so you only need memory for a single block.
How? We explicitly assign each block a role: to move the representation a little closer to the target than the block before it did. That role turns out to be precisely what a diffusion model does, step by step. Each block only needs to optimize its own objective and can be trained independently.
We validated this across five different architectures:
• ViT
• DiT
• Masked diffusion
• Autoregressive transformers
• Recurrent-depth transformers
In each case, performance is competitive with end-to-end training while using a fraction of the memory.
This perspective also extends naturally to recurrent-depth (Looped) transformers, which apply the same network iteratively and normally require expensive backpropagation through time (BPTT). Viewed through DiffusionBlocks, we can replace those multiple iterations with a single forward pass during training.
Read our paper and code, to learn more.
Paper: https://t.co/CRj96VGYQn
GitHub: https://t.co/eNW0K9Xh8E
🐟