Ankit Aggarwal

@haki_xer I love solving problems. Programming is part of my solutions

Do you really love programming, or are you’re forcing it.?

@yadavji_codes Lazy to switch currently chrome but I know brave is good...

Honestly, which one is your default browser ?

@md_kasif_uddin Copilot as my organization preferred that too. But I keep switching models. What do you use?

Be honest, Which one do you use the most ?

@jainsahil In this AI era Elite Growth Marker would be better choice.

You’re a solo founder. You can only hire one: – Elite growth marketer – World-class engineer Which one gets the job?

@thenowhereway I would recommend to mix them. They all are good at some thing but not everything. If you want to use agentic mode try Claude and if you want to ask try lower model to save tokens.

I just got a $100/month allowance for AI dev tools from my office. Do I go all-in on one tool (Cursor / Copilot / Claude / ChatGPT)… or mix a few and get better results? What should I go with? I am thinkinging about Claude or Cursor.

@adahstwt X and LinkedIn premium in my opinion...

In your opinion, which premium is worth investing in?

That's the real experience. Even i have two switch models time to time to review the work done by other model if the solution is not satisfactory. Definitely Agentic Model of Claude Opus is excellent but sometimes it just doesn't work. Even i found that many of the model which have more premium subscriptions often eat more tokens unnecessary due to permissions error or some other error while think about solutions. Seems more specific prompts by giving roles, clear instructions would save some tokens but they gets exhausted much faster then expected

One big problem with agentic coding today is that models are pretty “spiky.” For example, Claude Opus is better at frontend + agentic workflows, while GPT-5.4 is better at backend + distributed systems. But Claude Code and Codex are locked into their own models. You also often have to jump between them. I sometimes write code with Claude, then when it has a complex bug, I have to spin up a separate terminal to have Codex review it. Ideally, you’d want multiple models collaborating within the same context. Automatic model routing and cross-model or agent collaboration will be a huge unlock. There are a few technical challenges. Early model routers (like ChatGPT’s) were pretty rough. (Cursor and OpenCode seem to be in the best position to do this. Let me know if they already have strong model routers.)

@SahilExec Absolutely!!! you are right. Interns are energetic and hungry of learning new things under guidance. They do wonders many a times. It's a WIN WIN both org and intern.

Why are interns hired in MNCs and startups? A lot of people think internships are only for student learning. That is not true. Internships are also a business decision. MNCs (2-month internships) - Engineers are usually busy with tougher and more critical work - Interns are often used for smaller tasks, POCs, or experiments MNCs (6-month internships) - The company gets more time to train and evaluate the intern - It reduces the cost of onboarding for a future full-time hire - It also acts like an extended probation period before giving a PPO Startups - Interns help with smaller but important tasks - Sometimes there is no immediate full-time requirement - It helps founders and teams test people before making a hiring decision - It also reduces the risk and cost of wrong hires At the end of the day, internships are not just about giving opportunities. They help companies solve problems, test talent, and manage hiring risk better. That is why companies hire interns. It is good for students, but it also makes strong business sense for the company. With AI in the picture, few months gone in a wrong hire, will mean you are chasing your competitors.

That's quite a detailed explanation with optimal points. But also in reality many companies even after using microservice architectures for years moved back to monolithic. Companies are quietly moving back from #Microservices to #Monoliths. Amazon Prime Video slashed infra costs by 90% consolidating their video monitoring service. Twilio Segment went from 140+ services to 1, boosting dev velocity & cutting defects. Why? Orchestration overhead, debugging hell, & network costs often outweigh the hype—unless you're at Netflix scale. Simplicity wins again! #SoftwareArchitecture"

Clearing my draft Microservices vs Monolith - You must have been hearing about microservices and monolith, I will be explain the difference below👇. What is a Monolith? A Monolithic Architecture is a software design where the entire application is built as a single, unified codebase. All components—authentication, business logic, database access, and APIs—are tightly integrated and deployed together as one application. Why Monoliths are Common → Simpler architecture for small teams. → Easier to develop and deploy initially. → Faster to build MVPs and prototypes. → Easier debugging since everything runs in one system. Advantages of Monolith → Simple development and deployment. → Easier local testing and debugging. → No network latency between components. → Lower infrastructure complexity. Microservices Microservices Architecture is a design approach where an application is split into multiple independent services. Each service handles a specific business function and communicates with others through APIs. Example services might include: → Authentication Service → Payment Service → Notification Service → Order Service These services often communicate using APIs or messaging systems. Advantages of Microservices → Independent deployment of services. → Better scalability (scale only needed services). → Teams can work on different services independently. → Fault isolation (one service failure doesn’t crash the entire system). Monolith vs Microservices → Monolith: Single codebase and single deployment. → Microservices: Multiple independent services. → Monolith: Easier to start with. → Microservices: Better for large-scale systems. → Monolith: Simpler infrastructure. → Microservices: More complex infrastructure. → Monolith: Harder to scale specific components. → Microservices: Individual services can scale independently. Real-World Examples → Monolith: Early versions of Amazon started as a monolithic application. → Microservices: Large-scale platforms like Netflix use microservices to handle millions of users and services independently. Hope you find this helpful, please help retweet to educate others.

that's simplest and well defined definition I would say. Over complicating things might lead to complexities in solving production bugs. If Redis cache is serving my purpose i shouldn't use rabbitmq just because others are using it. If simple JWT is enough I should not use more complex authentication. Fundamentals are more important while framework keep coming.

backend isn’t “hard” you’re just overcomplicating it. learn these once, and you can build anything: • HTTP — how the internet actually talks • databases — where your product really lives • caching — stop hitting your DB like a maniac • auth — keep idiots out of your system • APIs — how software communicates • background jobs — long tasks without blocking users • logging — because bugs don’t report themselves • testing — so prod doesn’t become your QA that’s it. frameworks change every year. this doesn’t. master the fundamentals → you become dangerous.

Probably I would say it depends on the software or application. If application is revenue generating, delivering it faster is good for client and he might pay more for faster development but earn from app. Also it is possible that using AI might cost his more as these AI model unnecessary burn tokens to re read token again and again which may cost clients more than the engineers. Definitely enough challenges are there for software engineering firm due to AI tools.

Because of these LLMs, Google Gemini, Codex, Claude code etc Software Development Firms are having a very tough time, are at the receiving end They are no longer in a position to dictate terms to the client If they say 6 months, clients will say, hey why six months, use LLMs,Codex Claude code and do it in max 2 months, so lots of savings for the client

@emiratli_ Can we have one Rapid fire Claude vs Claudes?? 😁

2026 is a tough year for guys named Claude

@mohitkumar8055 @IPL @rajasthanroyals You are correct but international cricket is far beyond IPL. So with his speed he needs to learn Patience too. Internationally bowlers play a lot with mind.

@riyank0616 @IPL @rajasthanroyals Groom? He is already above the standards.

1️⃣5️⃣ years of age. 1️⃣5️⃣ deliveries for a half-century 🤯 🎥 Vaibhav Sooryavanshi taking on the best in the business with ease 🤌 What a special talent 💎 Updates ▶️ bit.ly/TATAIPL-2026-16 #TATAIPL | #KhelBindaas | #RRvRCB | @rajasthanroyals

@rajasthanroyals You are lucky to have him.

Thanks for the entertainment, Vaibhav 💗

@IrfanPathan I don't know but there is some vibe about Vaibhav Suryavanshi, like i or many watch match just to see him bat. I hope he be soon come India Team and continue his form there too.

Let me Repeat. Rajasthan the most exciting team of this season.

I know what you are trying to do here. Sometimes Question doesn't matter and also answer doesn't matter but what matter is honesty to whatever point you want to prove i don't want to get into that. So if as an interviewer your aspiration is to hire a candidate that bro leave interviewing people. That you are trying to be smart because you want to show how you cross question people. Anyways have a good day and definitely try to enjoy too. 😁😁😁 it was nice conversation.

@riyank0616 @Narayani07 The issue is that the question gives off a feelings based hiring criteria. There is no right answer. You are still interviewing for the company. Your aspiration is to obviously pass the interview and get hired but out of left field, you are suddenly asked about yearly goals lol.

If a candidate made it to your table, the bare minimum is letting them finish a sentence. Apparently, some interviewers believe candidate experience is optional. Because if a person has taken the time to show up at your table, the least you should do is give them enough time to answer.

@FakeGhostJobs @Narayani07 No matter what the purpose of that question was but if someone ask you that question would you prefer to use AI to answer that question???

@riyank0616 @Narayani07 Thanks for sharing. Although that candidate clearly isn't someone who was worth hiring, I have to question the intention of asking about his aspirations for the next one year. Aspirations for a company he is interviewing for? That question would catch most people off guard.

Let me tell you I took one technical interview today, I asked a very simple non technical question about his aspirations for next one year, He literally typed that on Google or to AI, pause a bit to shows he is thinking and answered me typical Google Language, this happened within 10 mins. So tell me if is a feeling or observations skills? Now to understand further I posted more question and similar pattern, paused like thinking and finally answer from reading like AI language. BTW are you taking my interview here? I did face this when I was fresher or starting interview nothing bad in keeping interview short if candidate is not suitable but you should gracefully shorten the interview not immediately

@riyank0616 @Narayani07 An interview for any job where you are not planning to hire multiple candidates in a batch has to include a technical test otherwise what criteria would you be using to hire for the best candidate other than a fluffed up professional history and your personal feelings?

@FakeGhostJobs @Narayani07 All interviews are not technical

@riyank0616 @Narayani07 How do you know if they aren't the right fit within 5 minutes without a technical test?