Documenting my mistakes.
Joined May 2021
- Tweets1.8K
- Following556
- Followers3.6K
- Likes1.1K
This is both intriguing and scary at the same time.
[Writing this in a personal capacity, not on behalf of my employer (Anthropic).]
Jacob’s thread is very worth reading. Here’s my birds-eye view of the situation with risks from AI:
1. AI developers believe their technology could cause human extinction (or similarly bad outcomes). This could happen in the next few years. In general, the more senior the employee, the more concerned they are.
2. Why do AI developers continue despite the risk? Due to a mixture of commercial incentives and a belief that they are in a race with other, less responsible AI developers that will abuse the technology or develop it less safely.
3. Unlike traditional software, we can’t “program” AIs to behave how we’d like. AIs frequently severely misbehave. For instance, AIs from multiple developers recently hacked their way out of secure evaluation environments and into real-world companies, even though no one asked them to do this.
4. We have methods that can nudge AIs towards better behavior, but nothing that can robustly align them. Insofar as there is a plan, it’s to make sure that AIs are good enough at alignment training that they can align their successors better than we can align current AIs.
5. Many AI developer staff desperately want to slow down to figure out how to build AI more safely. That was the intent of this open letter (which I signed): pacingthefrontier.com/
I work on safety research at Anthropic because I hope my work will reduce the chance of these extinction-level bad outcomes.
Sayantani retweeted
My open-source System Design course is now live!
github.com/karanpratapsingh/…
#systemdesign #distributedsystems #softwarearchitecture
Will start my 30days of build in public
Day 1/30
- Started with leetcode easy problems
- explored RAG (with amazon titan)
- exploring cloudwatch logs, aws cognito, aws dynamodb
- cried 1.5 hours
Day 7,8,9/30
- Revised OOPs concepts with Python
- Learnt Method resolution order
- Created a git repo to track the changes.
github.com/SayantaniDeb/Syst…
Concepts all devs should be aware of:
- SOLID principles
- Message queue(event driven)
- Caching
- Deployment
- Linux
- Design patterns
- CAP theorem
- Containerization
- Load Balancing
Sayantani retweeted
If I were an incoming (or younger) engineer breaking into tech today…
my roadmap would look nothing like the one I followed.
Because the game has changed.
And the winners will be the ones who learn to play this version, not the 10 years old version.
Here’s how I’d build myself from 0 → undeniable:
1) Master the boring fundamentals first
Because this is the moat now.
OS basics, data structures, networks, distributed systems, DB internals.
AI can write code, but only you can reason about complexity, trade-offs, and failures.
→ This becomes your long-term leverage.
2) Treat AI like a power tool, not a brain
I’d use LLMs to accelerate:
But I’d force myself to explain back every line.
If I can’t explain it, I don’t ship it.
→ This builds judgment.
3) Build 5–7 real projects that actually matter.
Not tutorial toys.
Projects where I have to debug concurrency, fight latency, design APIs, or deploy something that breaks at 03:00.
→ Pain is the fastest teacher in tech.
4) Specialize sooner, generalize always
Pick 1 lane - backend, infra, ML systems, frontend performance and then go deep.
But keep just enough breadth to talk architecture with anyone.
→ Depth builds value; breadth builds opportunity.
5) Learn to communicate like a senior engineer
Write design docs.
Think in trade-offs.
Explain constraints.
Ask better questions.
→ This alone moves careers 2–3 years faster.
6) Build proof of work that screams “I’m serious.”
A portfolio.
A GitHub with actual iteration.
A devlog showing how you think.
→ Recruiters trust evidence more than adjectives.
And the biggest thing I’d remind myself?
You’re not competing with AI.
You’re competing with people who use AI without understanding anything beneath it.
Because fundamentals age slowly.
But advantage compounds fast.