AI Thoughtmakers

Token costs are going down. Total AI spend is going up. And most people are only looking at half the picture.

In this episode of AI Thoughtmakers, we sit down with Saurabh Srivastav, Senior Software Engineer at GeekyAnts, for a sharp, no-fluff conversation about the real economics of AI — what it actually costs, where founders should and shouldn't be building, and why judgment is about to become the scarcest skill in the entire industry.

From why you should never build a generic AI app, to why Friday deploys are a death wish in the AI era, to why context engineering will outlive prompt engineering by a mile — this is the episode for engineers, founders, and leaders who want to understand the AI economy from the inside.

Key topics covered:

• The real cost of AI — what subscribers pay vs what AI companies actually spend

• Why token costs going down doesn't mean total AI spend is going down

• Where founders without big capital backing should and should not be building

• Why generic AI apps are a trap — and what to build instead

• Why judgment will be the scarce skill as implementation becomes easier

• One judgment engineer vs five prompt engineers — which one to hire and why

• Why context engineering matters more than prompt engineering going forward

• Why AI products fail after deployment — the verification problem nobody talks about

• Adding AI to an existing product — real innovation or just marketing?

• Customer support as the perfect example of AI integration done right

• Why deploying AI-powered software on a Friday can be a bad idea

Whether you're a founder trying to figure out where to place your bets, an engineer wondering where your value lies, or a hiring manager rethinking what skills actually matter — this episode gives you a clearer picture of where the AI economy is actually heading.

Subscribe for more conversations on AI, engineering, product development, and the future of software.

Connect with Saurabh:   / saurabhsrivastava108   

Connect with Prem:   /premgoswami  

AI is writing code, fixing bugs, and deploying software. So what does the future of software engineering look like when AI can handle more of the work?

In this episode of AI Thoughtmakers, we sit down with Jaspreet, Tech Lead at GeekyAnts and a team of one shipping and owning everything on his project, for a conversation that every engineer needs to hear right now — about what it actually means to be indispensable in an AI-native world, and why the engineers who are thriving aren't the ones who are scared of AI, but the ones who've made it their multiplier.

From why coding has always been just 20-30% of the job (AI just made it visible), to why you can outsource your thinking but never your understanding — this is the most grounded, practical take on the future of human engineering you'll find.

Key topics covered:

• Why there are two kinds of engineers today — those panicking and those shipping faster than ever

• The headlines that are misleading engineers — and what Google's "25% AI-written code" actually means

• What system-level thinking is — and why it's the only thing that will keep engineers relevant

• What you should never delegate to AI — customer calls, understanding, and judgment

• Why coding has always been 20-30% of the job — and AI just made that visible

• Why universities should stop treating syntax as mastery — and focus on fundamentals instead

• The one skill every engineer must master: debugging

• Why communication matters more than coding in the AI era

• How prompt engineering evolved into context engineering — and why that's actually bigger

• How managers need to evolve — from counting tickets to measuring system stability and outcomes

If you're an engineer anxious about where things are going, a tech lead building AI-native teams, or a student trying to figure out what skills actually matter — this episode will give you more clarity than any headline.

Subscribe for more conversations on AI, engineering, product development, and the future of software.

Connect with Jaspreet: LinkedIn -   / jaspreets0026   

Connect with Prem: LinkedIn -   / premgoswami  

AI can build your MVP before lunch. That doesn't mean it's ready for your users.

In this episode of AI Thoughtmakers, we sit down with Vishal, Tech Lead at GeekyAnts, and Roshan, Senior Software Engineer at GeekyAnts, for a conversation every non-technical founder needs to hear about what AI-generated code actually looks like under the hood, why it passes the demo but fails in production, and why the biggest misconception in the founder world right now is thinking you don't need engineers.

From exposed API keys hiding in plain sight to why AI will hallucinate its way through your codebase when you're not looking, this is the honest, no-filter guide to evaluating AI-generated code quality before it costs you everything.

Key topics covered:
• Why AI is powerful for MVP development but risky without an engineering layer
• The biggest founder misconception right now and why engineers are still non-negotiable
• Should non-technical founders trust AI-generated code? The answer is clear
• The most common warning signs in AI-generated code founders miss entirely
• What to verify before you deploy — compliance, security, and edge cases
• Why AI still can't handle long-term maintainability and code structure
• Senior software engineer vs. premium AI coding model — which one would you choose?
• Security risks such as exposed API keys, weak authentication, and PII leakage
• Can AI-generated code pass a compliance audit? What needs to happen first
• Who is responsible when AI-generated code causes a security breach?
• How software teams will look 5 years from now — and why engineers aren't going anywhere

If you're a non-technical founder building an app with AI, a startup leader evaluating AI-generated software, or an engineer explaining why an AI-built MVP is not ready for production, this episode will help you understand what happens between an impressive demo and reliable production software.

Subscribe for more conversations on AI, engineering, product development, and the future of software.

Connect with Vishal:   / vishalsh2299 
Connect with Roshan:   / roshan-kumar-ojha 
Connect with Prem:   / premgoswami  

Your architecture isn't broken. But the assumptions baked into it probably are.

In this episode of AI Thoughtmakers, we sit down with Aditya Prakash, Lead DevOps Engineer at GeekyAnts, for a deeply technical and refreshingly honest conversation about why traditional cloud infrastructure, microservices, REST APIs, and auto-scaling strategies struggle to support modern AI applications.

From why AI blows up your cloud bill before your CPU alert even fires, to why prompt injection is the new SQL injection, to why the one architecture pattern every team needs right now is an AI gateway, this is the episode your DevOps and platform engineering teams need to hear.

Key topics covered:

• What traditional software architecture actually means and why AI is breaking its core assumptions

• How AI infrastructure differs from traditional cloud infrastructure

• Why AI scaling is fundamentally different from CPU and RAM-based auto scaling

• How AI sends your cloud bill through the roof before you even realize it

• Why service communication and gateway timeouts break when AI enters the picture

• AI security best practices: Prompt Injection, PII protection, and AI governance

• Why auditing AI applications is harder than anything we've dealt with in traditional systems

• The Reverse Information Paradox and AI data privacy concerns

• Why every team needs an AI gateway and what it actually does

• How platform engineering needs to evolve to support AI infra properly

• GPU fundamentals every DevOps engineer should understand

If you're an engineering leader, DevOps engineer, or CTO trying to figure out how to make your infrastructure AI-ready without burning your budget or your security posture, this is the conversation that brings it all into focus.

Subscribe for more conversations on AI, engineering, product development, and the future of software.

Connect with Adithya: LinkedIn -   / aditya-2811   

Connect with Prem: LinkedIn -   / premgoswami   

Does hiring more software developers actually build better products?

In this episode of AI Thoughtmakers, we sit down with Rahul, Data Engineer and YouTuber (Start Practicing – 280K subscribers), and Arfin, Cloud Solution Architect at Microsoft, for one of the most honest conversations we've had on the podcast about why the companies hiring more engineers are solving the wrong problem, and why product thinking is the real competitive advantage in the AI era.

From why Apple, Airbnb, and Netflix don't just ship features to why AI is now helping companies build the wrong product faster and what it actually means to own your decisions when AI is writing the code, this episode will change how you think about what you're building and why.

Key topics covered:

Why hiring more developers is a capacity solution to a clarity problem

How big teams create communication overhead and still miss deadlines

What product thinking actually means and why it isn't just a PM's job

Finding clarity from ambiguity: the real skill is deciding what NOT to build

What separates Apple, Airbnb, and Netflix from companies that ship and fail

Has AI exposed weak product strategies or just helped companies build wrong products faster?

Why you should never outsource your decision making to AI

Startups vs enterprises: agility vs maturity. Who has better product thinking?

If you had a budget for just one hire — product thinker or developer?

Why the future belongs to people who are better decision makers

Whether you're a founder, product leader, software engineer, or someone building in the AI era, this is the episode that brings clarity back to the conversation.

Subscribe for more conversations on AI, engineering, product development, and the future of software.

Connect with Rahul: LinkedIn -   / rahul-khemka-aa487a10   

Connect with Arfin: LinkedIn -   / arfin-parween   

Arfin’s YouTube channel -    / @startpracticing   

Connect with Prem: LinkedIn -   / premgoswami   

Full-time hiring takes 3 to 4 months. The market expects results in the same timeframe. Something has to give.

In this episode of AI Thoughtmakers, we sit down with Vikas, Full Stack Engineer and Team Lead at GeekyAnts, for a no-jargon conversation about fractional engineering teams — what they actually are, when they make sense, when they don't, and why the fastest growing companies are already using them to ship faster without bloating their headcount.

From the house renovation analogy that finally makes fractional hiring click, to the biggest myth founders believe about handing over ownership, this episode cuts through the confusion around fractional hiring, outsourcing, and freelancing — and gives you a clear framework for when to use each.

Key topics covered:

• Why traditional tech hiring can slow down product development

• Fractional engineering teams explained in plain language — no jargon

• Fractional vs permanent hiring: speed vs longevity — and how to decide

• The biggest myth about fractional teams that's costing companies trust and output

• When you should NOT hire a fractional engineering team

• How knowledge transfer actually works — and why most companies get it wrong

• The real ROI of fractional engineering — beyond simply reducing costs

• How AI is changing the fractional model and shrinking the team sizes needed to ship

• Why the future of engineering is smaller teams shipping more products

Whether you're a founder building a product, trying to move fast without overhiring, an engineering leader plugging gaps in your team, or an enterprise looking to scale specific capabilities quickly — this episode gives you the playbook.

Subscribe for more conversations on AI, engineering, product development, and the future of software.

Connect with Vikas: LinkedIn -   / vikaskumar75   

Connect with Prem: LinkedIn -   / premgoswami   

AI is producing more code than ever. So why does it feel like quality is getting worse?

In this episode of AI Thoughtmakers, we sit down with Ankit and Joydeep, Tech Leads at GeekyAnts, for a no-filter conversation about what code quality and engineering excellence actually mean in the age of AI and why the rules of the game haven't changed as much as people think.

From who owns a $10 million bug when AI wrote the code, to why chasing 100% test coverage is overrated, to what engineering excellence will look like when AI is writing 95% of the code by 2030, this is the conversation senior engineers are having internally but rarely say out loud.

Key topics covered:

• Why more AI-generated code means more bugs — and why reviews are now more critical than ever

• How engineering priorities have shifted from writing code to questioning architecture

• The 100% test coverage myth — and what actually guarantees software quality

• Why business alignment matters more than advanced tech choices for startups

• Who is accountable when AI-generated code ships a bug to production?

• Why AI code needs more scrutiny than human-written code — and how to review it right

• The right way to do code reviews — stop checking variable names, start checking logic

• Why cutting corners is sometimes valid — but only if you document it

• What engineering excellence means in 2030 when AI writes 95% of the code

• Why the future engineer's most valuable skill is decision-making, not coding

If you're an engineer, tech lead, or engineering manager trying to stay ahead of the curve without losing sight of what actually makes software good — this episode is for you.

Subscribe for more conversations on AI, engineering, product development, and the future of software.

Connect with Ankit: LinkedIn -   / spanion   

Connect with Joydeep: LinkedIn -   / joydeep-nath-09b8a0164   

Connect with Prem: LinkedIn -   / premgoswami   

AI can code. AI can debug. AI can generate tests. So what exactly are developers supposed to be doing now?

In Episode 14 of AI Thoughtmakers by GeekyAnts, we sit down with Anamika and Bharat for a candid, no-filter conversation about what's actually changing on the ground for software developers in the AI era — and why the software companies that will win aren't the ones with the most AI tools, but the ones who know how to use them right.

From the Claude vs ChatGPT debate to what happens when AI goes down for a day, this episode gets into the real, unfiltered developer experience of working alongside AI in 2026.

Key topics covered:

• Are companies hiring developers or problem solvers — and what's the difference?

• Why Claude is becoming the go-to tool for architecture and coding over other AI platforms

• The risk of copying AI-generated code without understanding it

• How junior developers can grow in an AI-native development environment

• Why mentorship matters more when AI tools become part of daily work

• What happens to developers when AI tools go down for a day

• The growing dependency on AI — and whether that's a problem

• Why knowing your own codebase is still the most important skill a developer can have

• What software engineering will look like in 2030

• One thing every engineer should stop doing today — and one thing they should start

If you're a developer navigating the AI era, a tech lead mentoring the next generation, or a founder wondering what kind of talent you actually need — this conversation will give you a lot to think about.

Subscribe for more conversations on AI, software engineering, product development, developer productivity, AI-native development, and the future of software.

Connect with Anamika: LinkedIn -   / anamika-228824208   

Connect with Bharat: LinkedIn -   / bharat-dhiman-4855711a1   

Connect with Prem: LinkedIn -   / premgoswami   

Can AI really replace 70% of software development

? And if it can, what are developers actually supposed to do now?

In this episode of AI Thoughtmakers, we sit down with Sarika Gautham, Principal Technical Consultant at GeekyAnts, for a grounded, no-hype conversation about what AI is genuinely changing in product development and what it absolutely isn't.

From the real cost of token consumption to why a working prototype is not a production-ready product, Sarika cuts through the noise and gives digital platform founders, developers, and digital product leaders a clear-eyed picture of where AI creates opportunity and where it still has hard limits.

Key topics covered:

• Why AI can replace repetitive tasks but not human thought process and creativity

• How developers are evolving from coders to system designers and architects

• The token cost problem — why AI-powered development isn't as cheap as it looks

• Why founders should not stop hiring junior developers

• Will AI disrupt software outsourcing and IT services

• How AI has compressed the ideation-to-prototype journey for startups

• The biggest misconceptions founders have about AI-powered app development

• Why getting excited about a prototype is the biggest mistake founders are making

• Bold prediction: what product development will look like in 2030

• The most valuable skill to learn today — and it's not what most people expect

If you're a founder, developer, or product leader trying to figure out how to build smarter in the AI era, this is the conversation that brings expectations back to reality.

Subscribe for more conversations on AI, engineering, product development, and the future of software.

Connect with Sarika: LinkedIn -   / sarika-gautam-047749238   

Connect with Prem: LinkedIn -   / premgoswami 

No one can say traditional hiring will disappear, but there is a clear shift toward pods, contractors, and smaller AI-enabled teams.

In this episode of AI ThoughtMakers, Prem sits down with Suresh, Solution Architect at GeekyAnts, to explore one of the biggest shifts happening in technology teams today: the move from traditional full-time engineering teams to expert pods.

They discuss:

• What an AI-native pod is and how it works

• How pods operate differently from traditional outsourcing

• Why startups and enterprises are embracing pod-based execution

• The evolving role of engineering managers in the age of AI

• How AI is changing hiring, team structures, and productivity

• When to choose AI-native pods vs. full-time engineering teams

• Situations where building an in-house team still makes more sense

• How founders and CEOs can evaluate pod performance in the first 90 days

• The risks, limitations, and decision-making boundaries of pod-based models

• Whether the future of software development is shifting away from traditional hiring

Subscribe for more conversations on AI, product engineering, software development, and the future of work.

Connect with the Speakers

LinkedIn -Suresh

Connect with Prem: 

LinkedIn -   / premgoswami