Joined August 2009
- Tweets 1,009
- Following 241
- Followers 154
- Likes 57,100
44 Photos and videos
The biggest gains from Claude Code often come from improving the environment, not changing the model.
The main tweaks boil down to a common pattern:
• better project context
• clearer instructions
• tighter tool permissions
• reusable workflows
• less manual correction
1
27
The insight:
AI coding performance is increasingly a systems engineering problem.
A well-configured Claude Code setup can outperform a default setup using the exact same model because the agent has:
👉 better context
👉 better constraints
👉 better feedback loops
1
12
We're moving from prompt engineering to agent environment engineering. ⚙️🤖 #ClaudeCode #AIAgents #AIEngineering #SystemsEngineering
1
Microsoft’s Planner Agent is now generally available in Microsoft 365 Copilot.
What makes it interesting isn’t task management—it’s the shift toward agentic work management.
Instead of manually juggling Planner, To Do, Outlook, Teams, and meetings, the agent can:
1
16
The bigger trend:
We’re moving from productivity tools that store work → to agents that understand and coordinate work.
1
Project management is becoming less about managing tasks and more about managing intelligent workflows. ⚙️🤖 #Microsoft365Copilot #AIAgents #FutureOfWork
srikrishna acharya retweeted
Jun 16
Can anyone explain why I have to pay tax on dividends, interest income and capital gains when I invested with money which was already taxed?
340
356
3,053
139,379
Some religious rituals bring a sense of calmness.. and it is nothing to do with religion or god but just the process of all elements (fire, air, water, earth, and ether) chants make the space vibrant
4
Everyone is talking about AI Agents.
Very few are talking about Systems Engineering for AI Agents.
I've started a repository to bridge that gap:
📚 Modern Systems Engineering
1
1
6
Because building reliable agentic systems requires more than prompts: • Requirements Engineering • Architecture • Interfaces • Verification & Validation • Observability • Governance • Human-in-the-Loop Design
1
7
Many of the challenges we're seeing in AI today were solved decades ago in aerospace, telecom, and complex systems engineering.
Curious:
What systems engineering concepts do you think AI engineers should learn next?
Repo github.com/srikrishna3118/mo…
Feedback and contributions welcome
1
9
No matter what age you are 15 - 60. Most men always feel they can do a better job than the person playing in the field. Irrespective of the sports they follow.. they always imagine some magical scenario that they are playing and people are praising.. but it is always other way..
1
6
A 100-word AIgenerated email consume roughly a bottle of water's worth of cooling infrastructure. The exact number changes over time, the bigger insight remains:
Every query has a hidden physical footprint:
• data centers
• GPUs
• electricity
• cooling systems
• water usage
1
5
We often think of AI as software.
But at scale, AI is really infrastructure.
The next frontier isn't just building smarter models.
It's building models that are
more efficient
less energy-intensive
less water-intensive
The future AI will be shaped by thermodynamics more than algs
45
Atlético Madrid's Stance: Atlético Madrid firmly rejected the offer, publicly attacking Real Madrid and stating that Julián Álvarez is not for sale
Current Status: Barcelona remains interested, believing the player wants the move, but a €100 million offer was insufficient.
1
13
Real Madrid currently has no plans to return with an improved bid following Atlético's aggressive response
Other Transfer Updates:
Vitinha and João Neves: The agent for both players confirmed they are not for sale and are considered crucial to Paris Saint-Germain
20
The Julián Álvarez Saga:
Real Madrid's Bid: Real Madrid made a club-record €150 million bid for Julián Álvarez after failing to pursue Michael Olise due to Bayern Munich's refusal to negotiate
1
32