DevOps @Accenture | Tech | AI | Humor Sharing DevOps & AI insights + daily hustle ✨ Collabs: DM or 📩 [email protected]
Pune
Joined February 2021
- Tweets22K
- Following1.8K
- Followers23.3K
- Likes65.6K
Pinned Tweet
A survey of 300 engineering leaders: their teams spend 9.8 hours a week writing code and 16.9 hours debugging it.
79% say AI agents write code significantly faster, but the overall release cycle is "no faster than before."
91% had test escapes, serious defects or badly optimised code reach production at least once in six months.
Grain of salt: mostly C/C++ teams, and a debugging company commissioned it. Writing got cheap. Finding out what's broken didn't.
Moving the finding out earlier looks like this: the coding agent runs the TestSprite CLI before handing anything back. It tests the live preview like a real user; on a failure it gets the step, screenshots and a root-cause hypothesis, fixes it and re-runs. The exit codes are CI-friendly, so the same check can gate merges.
You don't get faster at debugging. You just do less of it in production.
Where does most of your week go now?
Most people in India rely entirely on their job salary for money.
They don't even explore how to make money online or earn beyond their jobs.
It's not entirely their fault. Since childhood, we've been taught to study, get a job, and earn money.
Nandkishor retweeted
A survey of 300 engineering leaders: their teams spend 9.8 hours a week writing code and 16.9 hours debugging it.
79% say AI agents write code significantly faster, but the overall release cycle is "no faster than before."
91% had test escapes, serious defects or badly optimised code reach production at least once in six months.
Grain of salt: mostly C/C++ teams, and a debugging company commissioned it. Writing got cheap. Finding out what's broken didn't.
Moving the finding out earlier looks like this: the coding agent runs the TestSprite CLI before handing anything back. It tests the live preview like a real user; on a failure it gets the step, screenshots and a root-cause hypothesis, fixes it and re-runs. The exit codes are CI-friendly, so the same check can gate merges.
You don't get faster at debugging. You just do less of it in production.
Where does most of your week go now?
Finally, no need to sit in front of your laptop just to click approve.
Just assign the task and go for a walk. Claude will take care of the rest.
Nandkishor retweeted
AI agents could potentially bypass the very systems designed to keep them under control.
METR and Redwood Research spent 3 weeks testing OpenAI’s internal agent monitoring and uncovered serious gaps
> Agents could delay monitoring
> Exploits could hide hundreds of actions
> Codex agents could potentially escape sandbox restrictions
OpenAI says production vulnerabilities have been patched, but METR hasn’t verified the fixes yet.
AI safety claims mean little without independent verification.
This summer, we invited METR and Redwood Research to pilot an embedded evaluation of our internal monitoring system, providing deep access to our infrastructure over three weeks.
Their findings helped strengthen safeguards as we expanded the system and will help shape future research collaborations with METR and other third party evaluators.
Wait, we no longer need to switch between multiple apps to create content for X?
Congratulations, Kushal! You understand creators pain points better than anyone, having grown to 50K+ followers yourself.
Will try it and share my feedback!
🚀 PostDesk is LIVE!
I built a tool to solve a problem I face as a tech creator.
Write. Style. Plan. Design. Your social posts, all in one tab.
Built for creators and engineers who share tech content.
Less tool switching. More creating.
Try it out 👇
🔗 postdesk.techfusionist.com/
Would love your feedback! ❤️
Don't listen to random people saying, "It's easy to achieve 40+ LPA."
Everyone has their own journey. Nobody starts at the top. Most people start small and grow their packages over time.
Don't compare yourself with others. Focus on improving every day, upskilling, and switching strategically.
Those who've been through it know how difficult the Indian job market is. It's not a cakewalk. Freshers are still struggling to land their first job.
Trust your journey. Keep growing.
Everyone wants ₹40–50 LPA.
Very few want the journey that comes with it.
People earning ₹40–50L+ often made uncomfortable choices:
→ Risky switches
→ Moving cities
→ Longer hours
→ Weekend work
→ Taking career risks
Meanwhile, some people want ₹50L without accepting any uncertainty.
There’s nothing wrong with choosing stability.
But you can’t demand the upside without accepting the trade-offs.
Ambition without commitment eventually becomes frustration.
Nandkishor retweeted
Everyone wants ₹40–50 LPA.
Very few want the journey that comes with it.
People earning ₹40–50L+ often made uncomfortable choices:
→ Risky switches
→ Moving cities
→ Longer hours
→ Weekend work
→ Taking career risks
Meanwhile, some people want ₹50L without accepting any uncertainty.
There’s nothing wrong with choosing stability.
But you can’t demand the upside without accepting the trade-offs.
Ambition without commitment eventually becomes frustration.
Nandkishor retweeted
Indian VS Non-Indian interviewer, if you ever get a chance to choose between these two.
Choose the non-Indian one.
Here's why:
1. Better engagement
- Non-Indian interviewers often keep their cameras on and engage with you.
> while some Indian interviewers leave you talking to a blank screen.
2. A relaxed start
- They often begin with greetings and small talk to help you feel comfortable.
> Some Indian interviewers jump straight into “Introduce yourself.”
3. A fair evaluation
- Good interviewers focus on your skills and potential.
> Indian interviewers set unrealistic expectations or look for reasons to reject them.
Of course, not every interviewer is the same.
An interview should test your skills, not your ability to handle unnecessary pressure.
How is your experience between these two?
AI agents could potentially bypass the very systems designed to keep them under control.
METR and Redwood Research spent 3 weeks testing OpenAI’s internal agent monitoring and uncovered serious gaps
> Agents could delay monitoring
> Exploits could hide hundreds of actions
> Codex agents could potentially escape sandbox restrictions
OpenAI says production vulnerabilities have been patched, but METR hasn’t verified the fixes yet.
AI safety claims mean little without independent verification.
This summer, we invited METR and Redwood Research to pilot an embedded evaluation of our internal monitoring system, providing deep access to our infrastructure over three weeks.
Their findings helped strengthen safeguards as we expanded the system and will help shape future research collaborations with METR and other third party evaluators.
Nandkishor retweeted
This celebrity face-swap edit looks like it took hours, but the workflow is way simpler than you think 👀
- Open CapCut Web, go to Video Editor, click the icon in the top right, and copy the prompt.
- Paste it into ChatGPT and send it to install the CapCut plugin.
- Then bring it into Codex and let it help turn your creative idea and footage into an actual video.
- From rough cuts to transitions, captions, and music, CapCut Web helps bring the edit together.
Want to make your own celebrity face-swap video?
You bring the concept and source clips, while Codex helps execute your vision.
And the best part? You can still fine-tune the editable draft in CapCut Web.
#capcutweb #capcutchatgptplugin #capcutcodex #capcutmarketing
🤝 Paid partnership
Indian VS Non-Indian interviewer, if you ever get a chance to choose between these two.
Choose the non-Indian one.
Here's why:
1. Better engagement
- Non-Indian interviewers often keep their cameras on and engage with you.
> while some Indian interviewers leave you talking to a blank screen.
2. A relaxed start
- They often begin with greetings and small talk to help you feel comfortable.
> Some Indian interviewers jump straight into “Introduce yourself.”
3. A fair evaluation
- Good interviewers focus on your skills and potential.
> Indian interviewers set unrealistic expectations or look for reasons to reject them.
Of course, not every interviewer is the same.
An interview should test your skills, not your ability to handle unnecessary pressure.
How is your experience between these two?
Everyone wants ₹40–50 LPA.
Very few want the journey that comes with it.
People earning ₹40–50L+ often made uncomfortable choices:
→ Risky switches
→ Moving cities
→ Longer hours
→ Weekend work
→ Taking career risks
Meanwhile, some people want ₹50L without accepting any uncertainty.
There’s nothing wrong with choosing stability.
But you can’t demand the upside without accepting the trade-offs.
Ambition without commitment eventually becomes frustration.
Nandkishor retweeted
2026 is all about AI agents.
And now, AI agents can have their own email inboxes.
AgentMail lets AI agents send, receive, and act on emails just like humans.
If you also want to give your AI agent its own inbox.
BONUS: Comment "YOUGOTMAIL" on the launch post and get 1 month of the Startup plan for FREE.
🤝 Paid partnership
Today, @agentmail has processed over 1 Billion emails.
A year ago, we processed under 20 million.
So we figured it was time for AgentMail to grow up a little too :)
Nandkishor retweeted
OpenAI says GPT-5.6 Sol helped uncover a critical PHP vulnerability before attackers could exploit it. Wiz independently discovered and reported both flaws.
For DevOps engineers, this is a reminder to automate security scanning, patch vulnerabilities faster, secure CI/CD pipelines, and verify production deployments.
OpenAI models helped researchers—operating as part of our Patch the Planet initiative—discover critical security flaws in PHP, an open-source programming language used to run an estimated 70% of websites.
Over the past two months, we led a coalition of maintainers, researchers, and security providers to prepare fixes and protections ahead of disclosure. This duration was necessary because the fix was quite intricate and required a major change to the PHP engine. Every site using PHP can now safely upgrade their version to stay secure, in addition to applying mitigations from various security vendors.
Due to the severity of these flaws and PHP’s widespread use, we consider this bug to be one of the most impactful vulnerabilities in recent memory. It also marks a new phase, where AI can find security bugs with internet-scale impact.
This is an example of the Defender’s Window in action: to our knowledge, frontier AI was able to find this vulnerability ahead of attackers, giving defenders time to prepare.
openai.com/index/coordinated…
Readers added context they thought people might want to know
OpenAI may claim a role in the coordinated response and patching process. But the initial discovery and exploit development, however, occurred several months earlier by an independent researcher.
x.com/lcfr_hax/statu…
Nandkishor retweeted
After trying crypto, NFTs, freelancing, content creation, and a full-time IT job, one thing I’ve realised:
Your income is never guaranteed to stay the same. At some point, it will go down.