Instead of watching an hour of Netflix, watch this 2-hour Stanford lecture on AI careers. It will teach you more about winning in the AI race than all the AI content you’ve scrolled past this year.
This 2 hour Stanford lecture shows exactly how Stanford trains it's engineers to build AI systems. It's more practical than every Claude tutorial & prompting threads you've seen.
Bookmark & give it 2 hours, no matter what. It'll be the most productive thing you do this weekend.
Most people think using Claude Code is about writing better prompts.
It’s not.
The real unlock is structuring your repository so Claude can think like an engineer.
If your repo is messy, Claude behaves like a chatbot.
If your repo is structured, Claude behaves like a developer living inside your codebase.
Your project only needs 4 things:
• the why → what the system does
• the map → where things live
• the rules → what’s allowed / forbidden
• the workflows → how work gets done
I call this:
The Anatomy of a Claude Code Project 👇
━━━━━━━━━━━━━━━
1️⃣ CLAUDE.md = Repo Memory (Keep it Short)
This file is the north star for Claude.
Not a massive document.
Just three things:
• Purpose → why the system exists
• Repo map → how the project is structured
• Rules + commands → how Claude should operate
If CLAUDE.md becomes too long, the model starts missing critical signals.
Clarity beats size.
━━━━━━━━━━━━━━━
2️⃣ .claude/skills/ = Reusable Expert Modes
Stop repeating instructions in prompts.
Turn common workflows into reusable skills.
Examples:
• code review checklist
• refactoring playbook
• debugging workflow
• release procedures
Now Claude can switch into specialized modes instantly.
Result:
More consistent outputs across sessions and teammates.
━━━━━━━━━━━━━━━
3️⃣ .claude/hooks/ = Guardrails
Models forget.
Hooks don’t.
Use hooks for things that must always happen automatically.
Examples:
• run formatters after edits
• trigger tests after core changes
• block sensitive directories (auth, billing, migrations)
Hooks turn AI workflows into reliable engineering systems.
━━━━━━━━━━━━━━━
4️⃣ docs/ = Progressive Context
Don’t overload prompts with information.
Instead, let Claude navigate your documentation.
Examples:
• architecture overview
• ADRs (engineering decisions)
• operational runbooks
Claude doesn’t need everything in memory.
It just needs to know where truth lives.
━━━━━━━━━━━━━━━
5️⃣ Local CLAUDE.md for Critical Modules
Some areas of your system have hidden complexity.
Add local context files there.
Example:
src/auth/CLAUDE.md
src/persistence/CLAUDE.md
infra/CLAUDE.md
Now Claude understands the danger zones exactly when it works in them.
This dramatically reduces mistakes.
━━━━━━━━━━━━━━━
Here’s the shift most people miss:
Prompting is temporary.
Structure is permanent.
Once your repository is designed for AI:
Claude stops acting like a chatbot...
…and starts behaving like a project-native engineer. 🚀
The government wants AI doctors.
In 60 days, four new initiatives have launched:
- ARPA-H ADVOCATE: funding an autonomous AI cardiologist.
- CMS ACCESS: pays for outcomes, not visits, with reimbursement set that you have to use software to scale.
- FDA TEMPO: lets AI devices deploy before full clearance
- Utah: AI can prescribe medication (refills at first).
AI is coming to care delivery.
Apache Kafka 4.2.0 is about to drop any minute now! 🔥
A heavy release, it introduces close to 32,000 new lines of Java. What comes with all that code?
Here's everything you should know about it:
1/8 🧵
I'm giving Claude Code for FREE
(yes the world's best AI -- at no cost at all)
My mission is to get the next billion people to write great software, but so many are constrained by costs.
We solved this by combining AI with ads (only contextual and relevant ads, no spam).
In return, you get the best AI models without needing a credit card.
For the next batch, I will choose 500 people randomly who have done the following steps:
1. Comment "GIVE ME SANTA CLAUDE" below
2. Like and retweet this
3. Follow @NamanyayG and @supermemory
4. Sign up on the link in the next tweet
Watch this quick video to understand how it works!
We’re excited to announce that @UBS has successfully completed the world’s first in-production, end-to-end tokenized fund workflow leveraging the Chainlink Digital Transfer Agent (DTA) technical standard.
https://t.co/vY8LIKEPlI
UBS—one of the world’s largest private banks with over $6 trillion in AUM—worked with its in-house tokenization unit UBS Tokenize and DigiFT to showcase a live, in-production tokenized fund transaction that leverages the Chainlink DTA technical standard to complete the first-ever subscription and redemption request of a tokenized fund.
This development marks a major achievement that builds upon prior work between UBS and Chainlink within the Monetary Authority of Singapore’s Project Guardian initiative and proves how fund operations can be seamlessly automated onchain for increased efficiency and utility gains. In this live transaction, DigiFT functioned as the onchain fund distributor and leveraged the DTA standard to successfully request and process a subscription and redemption order.
This new end-to-end tokenized fund workflow can cover every stage of the fund lifecycle, including order taking, execution, settlement, and data synchronization across all onchain and offchain systems.
The DTA technical standard leverages key Chainlink platform capabilities, including:
• Chainlink Runtime Environment (CRE) for orchestration across onchain environments and existing in-house systems used by financial institutions.
• Cross-Chain Interoperability Protocol (CCIP) for interoperability across any public or private chain.
• Automated Compliance Engine (ACE) for programmable compliance.
• NAVLink for robust pricing inputs required for fund subscriptions and redemptions.
Our TPUs are headed to space!
Inspired by our history of moonshots, from quantum computing to autonomous driving, Project Suncatcher is exploring how we could one day build scalable ML compute systems in space, harnessing more of the sun’s power (which emits more power than 100 trillion times humanity’s total electricity production).
Like any moonshot, it’s going to require us to solve a lot of complex engineering challenges. Early research shows our Trillium-generation TPUs (our tensor processing units, purpose-built for AI) survived without damage when tested in a particle accelerator to simulate low-earth orbit levels of radiation. However, significant challenges still remain like thermal management and on-orbit system reliability.
More testing and breakthroughs will be needed as we count down to launch two prototype satellites with @planet by early 2027, our next milestone of many. Excited for us to be a part of all the innovation happening in (this) space!
@passportsevamea Sir for wife the address proof can be husband passport. This is part of the FAQ and the form. The AOP of Hubli Dharwad PSK it can't be. Can you please help in this (either update the document FAQ and drop down or ask AOP to consider it )
@passportsevamea What to do if Hubballi PSK AOP does not accept husband passport as the proof of address. This was harrassing experience as it is mentioned as valid document for wife address.
@passportsevamea Please note the Mobile OTP is not working. The OTP is not sent across. This is for renewal of passport to capture populate all the provided details.
BREAKING 🚨: NotebookLM is working on Slides generation! Soon, it will be able to generate a presentation from your sources.
It will be huge for Enterprise and Business customers!
Tons of SAAS apps build the same thing 👀
Massive. 🤯
Alibaba introduced a new GPU pooling system Aegaeon that makes AI model serving much more efficient.
Claims an 82% cut in Nvidia GPU use for serving LLMs by pooling compute across models.
In a 3+ month beta on Alibaba Cloud’s marketplace, H20 GPUs dropped from 1,192 to 213 while serving dozens of models up to 72B parameters.
The regular Cloud model hubs skew toward a few hot models, so many GPUs sit idle for cold models, and Alibaba measured 17.7% of GPUs handling only 1.35% of requests.
Aegaeon addresses this with token-level auto-scaling, which lets a GPU switch between models during generation instead of waiting for a full response to finish.
By slicing work at token boundaries and scheduling small bursts quickly, the system keeps memory warm and compute busy with little waste.
With Aegaeon, a single GPU supports up to 7 models versus 2 to 3 in other pooling systems, and switching latency drops by 97%.
Cold models load weights just in time when a request lands, then borrow a brief slice of compute without locking an entire GPU.
Hot models keep priority, so heavy traffic stays smooth while sporadic models borrow capacity in short bursts.
The wins apply to inference, not training, because generation happens token by token and fits fine-grained scheduling.
The timing suits China’s chip limits, where H20 targets inference workloads and domestic GPUs are ramping, so fewer chips can cover more traffic.
If Aegaeon generalizes, operators can lower cost per token, raise fleet utilization, and delay new GPU purchases without hurting latency for popular models.
Tradeoffs still exist, like uneven memory needs across models, long sequences that reduce preemption points, and scheduler overhead during traffic spikes.
----
scmp. com/business/article/3329450/alibaba-cloud-claims-slash-nvidia-gpu-use-82-new-pooling-system