Feb 2, 2024 · 55m · neon-show
Deepfakes, AGI, Jobs Under Threat & More | DevRev Founder, Manoj Agarwal | Neon Show
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this in-depth interview, DevRev co-founder Manoj Agarwal joins Siddharth Ahluwalia to discuss building AI-first enterprise architectures, the interface revolution democratizing software development, workforce automation, and the critical guardrails needed against deepfakes and AI hallucinations.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
Siddhartha as informed peer 3.9 Guest teaching 4.3 Guest disagreement 1.1 Siddhartha pushing back 1.7
speaking balance: gold is Siddhartha, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 49:36 Manoj challenges assumption of single right answers
When the host asserts ChatGPT renders search options redundant by giving the single right answer, Manoj directly refutes the premise, arguing humans still demand optionality and personal judgement.
Hardest push from Siddhartha ▶ 16:26 Host presses on autonomous bots breaking human rulesThe host rejects Manoj's safe assumption that bots will obey human rules, proposing a scenario where a goal-seeking bot overrides constraints because it possesses superior data.
Biggest teaching moment ▶ 3:03 Manoj breaks down NPS mathematical scoringManoj clarifies the common misconception confusing NPS with CSAT, explicitly explaining the negative scoring for detractors and zero weight for passive ratings.
Siddhartha holds their own ▶ 47:00 Host synthesizes the full evolution of UI paradigmsThe host demonstrates strong domain synthesis by charting how software interaction evolved from command line interfaces to button-heavy graphical UIs, and full circle back to conversational text.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Siddhartha as informed peer | Guest teaching | Guest disagreement | Siddhartha pushing back | Why |
|---|---|---|---|---|---|---|
| Episode Preview: Deepfakes, AGI, and Developer Explosion | 3 | 2 | 1 | 1 | The host opens the show and reviews Manoj's background at Nutanix. Manoj gently corrects the host on the exact year he joined Nutanix (2013 rather than 2009). | |
| Customer Obsession: Demystifying Net Promoter Score and Growth | 3 | 6 | 1 | 1 | The host asks about competing with Salesforce, prompting Manoj to educate the audience and host on the mathematical mechanics of Net Promoter Score (NPS) versus CSAT. | |
| The Genesis of DevRev: Reimagining Enterprise Data Models | 2 | 6 | 1 | 1 | The host asks Manoj to explain data models to a child. Manoj provides a structured breakdown of product, customers, and work tracking across Jira and Zendesk. | |
| Building AI-First: The Rise of Enterprise Copilots | 3 | 5 | 1 | 2 | The host asks for basic definitions of copilots and whether machines can truly think. Manoj explains bidirectional human-machine assistance with aviation and logistics analogies. | |
| The Reality of AGI, Safety Protocols, and Ethics | 4 | 4 | 2 | 4 | The host pushes a hypothetical scenario where an intelligent bot might override human safety rules to achieve a customer objective. Manoj counters that bots will remain bound by defined regulatory guardrails. | |
| ChatGPT's Watershed Moment and Large Language Models Explained | 3 | 6 | 1 | 1 | The host asks how to explain LLMs to a student. Manoj breaks down natural language processing, 175 billion parameter tuning, and how ChatGPT serves primarily as a UI interface layer over existing models. | |
| Deepfakes, Disinformation, and Geopolitical Threats in AI | 5 | 2 | 1 | 2 | The host shares examples of deepfakes including PM Modi's garba video and the evolution of phishing attacks. Manoj and the host mutually examine geopolitical risks and lack of international legal recourse. | |
| AI Workforce Disruption: Jobs at Risk and Automation | 5 | 3 | 1 | 1 | The host identifies vulnerable job categories like news anchors and support agents. Manoj supplements this with examples like SDRs, program managers, and local government initiatives to cut workforce costs. | |
| Inversion of Control: Prompting and a Billion Developers | 4 | 6 | 1 | 1 | Manoj introduces the concept of inversion of control where machines now learn human language instead of humans learning programming languages. The host adds observations on prompt engineering demand. | |
| Gen Z Adoption, Guardrails, and Reinforcement Learning | 3 | 5 | 1 | 1 | The host inquires about Gen Z adoption and reinforcement learning. Manoj uses personal observations of his daughters' search habits and explains reinforcement learning through human escalation and thumbs up/down signals. | |
| Evolution of the Indian Tech Ecosystem: Services to Products | 4 | 4 | 1 | 2 | The host asks why Manoj spends more time building in India now than during Nutanix. Manoj explains the macroeconomic shift from an IT services/maintenance mindset to primary product ownership and technical career tracks. | |
| Democratizing the AI Tech Stack and Modernizing Interfaces | 6 | 3 | 1 | 2 | Manoj compares AI APIs (LangChain, vector databases) to the democratization of AWS. The host provides a sharp synthesis tracking software UI from command prompts to dense button layouts back to natural language. | |
| Super Apps, Enterprise Guardrails, and Mitigating AI Hallucinations | 5 | 4 | 2 | 3 | The host questions the survival of individual apps in favor of super-interfaces and raises medical hallucination risks. Manoj explains that humans hallucinate constantly in organizations and shows how enterprise AI mitigates this via strict document boundaries. |
Statements from this episode (21)
Agarwal: Nutanix hit $1.6B software revenue by 2020 departure
“In 2020, when both Dheeras and I, we left the company, it was at 1.6 billion dollar in software revenue.”
Agarwal: Nutanix spent $1B on sales to generate $1.6B revenue
“Nutanix itself, like, 1.6 billion dollar that we were bringing revenue. We were spending close to four hundred million dollars just in R&D product side. Hundred million dollar in customer support, close to a billion dollar in sales and marketing, let's say.”
Agarwal: Humans will become copilots to autonomous enterprise machines
“The thing is that when machine has the data and if machine is able to do the work, can you let machine do it? And when machine doesn't know how to do that work, can it call the human to do the work? So in that sense, human can act like a copilot to machine.”
Ahluwalia: OpenAI's Team Is Divided on Whether to Achieve AGI
“And when they started OpenAI, the goal was to achieve AGI. And now everybody in the team is divided on whether we should achieve AGI or not.”
Ahluwalia: AI bots may start overriding human-designed rules within a year
“And probably in some time, not very long, like six months, one year, there can be a bot that can say, Hey, this rule designed by the human is not good enough. So my job is to give the best outcome to the customer. So, so I'll build my own rules or ignore these…”
Agarwal: ChatGPT succeeded by giving pre-existing LLMs a consumer interface
“And chart GPT was nothing but a UI representation or UI interface through which that human could go and interact with the LLMs or GPT models. That's what it provided a mechanism because prior to that, that, okay, through the APIs and you, by writing the code a…”
Ahluwalia: Phishing attacks will become smarter, targeting users over repeated attempts
“So I think now phishing attacks will become more intelligent. So they'll not try to fish on the first attempt. They'll try to fish on the hundred.”
Agarwal: Information retrieval jobs will be the first displaced by AI
“If you're not producing something yourself, and most of the time that, okay, a question is being asked to you, and you have to go and look up for that information, search for that information, or you read about that information. That's what you are going on an…”
Agarwal: Many sales development rep tasks can be automated by AI
“Even like the jobs that what in the sales development trip that they do, in fact, take the information and they just have to go and speak to the customers, like do a lot of dial and a lot of email and a lot of campaign and so on. Many of these things that can …”
Agarwal: AI regulation will create internal corporate compliance and ethics jobs
“The very first thing, like just the tons of regulations. That means that every company will require people. More regulators inside of ethical committee and compliance, and I mean, tons of jobs that will just right there will get created inside the company itse…”
Agarwal: AI brings an inversion of control where machines learn human language
“Inversion of control is previously for machine to be able to go and do the work. We had to go and write the, ah, language in which machine can understand it, which is like the programming languages that we had to go and write. That's the only way that machine …”
Agarwal: Prompt engineering will be automated by domain-specific bots
“My thinking is that even that will become a bot that will start to prompt based on the domain area that, that I am, that it's able to go and feed that information. So right now, obviously prompt engineering, because then you can get the answer the way that you…”
Agarwal: Gen Z Uses ChatGPT Over Google for Search
“Instead of going to Google to search for the information, they're searching on ChatGPT. They're just going and searching on the ChatGPT now, or asking the question and getting the more precise answer. And even for learning things that they're going more to thi…”
Agarwal: Most Enterprises Will Train Domain LLMs Over General Models
“So most enterprises that you will see that they will just do it that way. In many cases that you'll also see that especially for the enterprises again, that they will go and train the LLMs with the data set for their domains, with their data. Instead of just g…”
Agarwal: Indian tech satellite teams now build primary products over maintenance
“Every extension of the team or company that they were getting built here, they were doing maintenance work mostly. They are not building product. You'll see that the one dot of the product is built somewhere else. It's always like the, maybe enhancement, a sma…”
Agarwal: Indian engineers now choose technical tracks over management roles
“Previously people wanted to become manager very, very fast. So I have three years, five years of experience. Now I want to become manager, which was like the social pressure that the people had. And that also had changed dramatically. Now people wants to stay …”
Agarwal: Creating a billion developers requires AI-driven conversational interfaces
“When we say there are going to be a billion developers, let's say that's what Mr. Vinod Khosla is saying. And that simply means that how do you really go and create a billion developer? The only way that can happen is when you make the interface so simple, rig…”
Agarwal: Centralized conversational interfaces routing backend apps will arrive much sooner
“From the user perspective, they can just come and ask and it has to leave then transfer to the right app... You're right. That's like, okay, I can come to the Google. I can ask a question and it should just figure it out. Like, okay, which app and how to get t…”
Agarwal: Humans hallucinate just like AI when answering without precise knowledge
“When we talk about hallucination, even human hallucinate, right? Hallucination is like, okay, you don't know the answer, but you try to like relate it to something else and try to go and answer something. And which simply means that, okay, it's not a precise a…”
Agarwal: Enterprises will adopt AI only within strictly bounded datasets
“What like every enterprise, the way that they're going to go and adopt the AI is like, okay, here is this, you just, I want you to just play within this boundary. Here is the information that I'm providing you. If you can get the answer, then answer it. If not…”
Agarwal: AI is outperforming human workers across many law firm jobs
“We already see on the legal side that it's making quite a bit of an impact already. You can go and talk to the lawyers. They are already talking about, okay, there are so many jobs right now in the company that the machine is able to do much, much better job v…”