What's new: Nobel-winning economist Daron Acemoglu shared his updated perspective on AI's impact, particularly on agentic AI, emphasising augmentation over full job replacement.
Acemoglu's updated view challenges the widespread "AI jobs apocalypse" narrative, suggesting that while AI agents are advancing, their primary role will be to augment human work rather than replace entire jobs. He highlights the inherent complexity of human tasks, which often involve numerous different skills and coordination.
His perspective is rooted in economic analysis, which previously estimated a small productivity boost from AI, with no significant job displacement. This view counters the popular notion that AI agents, despite growing independence, can fully replicate the intricate orchestration of tasks performed by humans, such as an X-ray technician juggling 30 different tasks.
Acemoglu estimated in 2024 that AI would give only a small boost to US productivity and not obviate human work.
Studies repeatedly find that AI is not affecting employment rates or layoffs.
An X-ray technician juggles 30 different tasks.
Why it matters: This reframes the conversation around AI's impact on labour, pushing product managers to design AI solutions that enhance human capabilities rather than attempting full automation. It underscores the value of understanding complex human workflows and integrating AI as a supportive tool, potentially leading to more effective and socially beneficial products.
We're thinking: Acemoglu's argument for augmentation over replacement focuses on the immediate limitations of AI to handle the full scope of complex human jobs. However, this perspective may underestimate the potential for "job decomposition," in which AI chips away at individual tasks within a role until the human component becomes negligible, subtly shifting the definition of "augmentation" over time.
PM Interview Prep Club just launched structured courses.
Two are live. Three more launch in May covering Growth PM, Technical PM, and Design Thinking for PMs. Each course is built around real projects and the exact skills interviewers test, not theory. Early access members get the lowest price we'll ever offer.
What's new: MIT Technology Review published an article advocating for "customer-back engineering" as a strategy to foster breakthrough AI innovation by prioritising customer needs.
The article champions "customer-back engineering," an approach that starts with a deep understanding of customer problems and then works backwards to develop technology solutions, particularly in AI. This contrasts with a "technology-first" approach that often leads to solutions in search of problems.
This method aims to overcome the common pitfall of organisations failing to capture the expected value of digital investments due to a disconnect between technology development and actual user needs. By embedding engineers closer to customers, companies can foster "sideways innovation" and create more impactful, relevant AI products.
Organisations capture less than one-third of the value expected from digital investments, according to McKinsey research.
Ashish Agrawal, VP at Capital One, states, "When you get your engineers closer to customers, you get a lot more sideways innovation."
The customer-back engineering approach contrasts with a technology-first approach to innovation.
Why it matters: For AI product managers, this emphasises the need to prioritise user empathy and problem validation over technological novelty. It provides a robust framework to ensure AI products solve real-world problems effectively, driving higher adoption and value capture, rather than becoming sophisticated but unused tools.
We're thinking: The call for "customer-back engineering" in AI is crucial, yet it creates a structural tension with the "research-first" culture often prevalent in breakthrough AI labs. Bridging this gap requires product leaders to not only bring engineers closer to customers but also to actively translate nascent technological capabilities into potential customer value propositions, even before customers themselves know what to ask for.
What's new: Researchers introduced RateQuant, a method for optimal mixed-precision KV cache quantisation for LLMs that leverages rate-distortion theory to improve efficiency.
RateQuant addresses the significant memory bottleneck caused by the KV (Key-Value) cache in large language models during inference. This novel technique applies rate-distortion theory to quantise the KV cache, allowing for optimal compression while minimising information loss.
The KV cache grows linearly with sequence length, consuming substantial GPU memory during LLM inference, especially for long-context applications. RateQuant strategically reduces the precision of these cached values, specifically using a mixed-precision approach to maintain model quality while drastically cutting memory footprint and improving throughput.
KV cache grows linearly with sequence length, a primary memory bottleneck.
RateQuant reduces KIVI's perplexity from 49.3 to 14.9 (70% reduction) on Qwen3-8B at an average of 2.5 bits.
RateQuant adds zero overhead at inference time.
Why it matters: This breakthrough directly impacts the economic viability and scalability of deploying LLMs, especially for real-time and long-context applications. PMs building AI products can anticipate lower serving costs, enabling more complex features and broader accessibility, while also potentially supporting a new generation of LLMs with even larger context windows.
We're thinking: RateQuant's zero-inference overhead and significant perplexity reduction will force a re-evaluation of current LLM-serving infrastructure roadmaps. This makes mixed-precision KV cache quantisation a standard and expected feature within the next 12-18 months, rather than a specialised optimisation for a few leading-edge labs.
What's new: OpenAI reported that ChatGPT adoption surged in Q1 2026, with the fastest growth among users over 35 and more balanced gender usage.
OpenAI's latest update indicates a significant expansion of ChatGPT's user base in Q1 2026. The growth was particularly notable among older demographics, specifically users over 35, and also showed a move towards more equitable gender distribution in its user metrics.
This broadening adoption signals a shift beyond early tech-savvy adopters to a more mainstream audience. The increasing accessibility, improved usability, and growing awareness of AI's practical applications are likely contributing factors, moving ChatGPT from a niche tool to a widely adopted platform for various tasks.
ChatGPT adoption surged in Q1 2026.
Fastest growth was among users over 35.
Usage saw a more balanced gender distribution.
Why it matters: This data is critical for product managers developing AI applications, as it confirms the mainstreaming of generative AI. Understanding these demographic shifts allows teams to tailor features, marketing, and onboarding strategies to resonate with a broader, more diverse audience, moving beyond initial power users to capture significant market share.
We're thinking: The broadening adoption of ChatGPT among demographics in Q1 2026 confirms that generative AI is crossing the chasm from early adopters to the early majority, following a classic technology adoption lifecycle. This shift means product strategies for AI need to evolve from focusing on novel capabilities to emphasising reliability, integration into existing workflows, and addressing diverse, practical use cases for a mainstream audience.
AI as Augmentation, Not Replacement: The economic perspective from Acemoglu suggests PMs should focus on designing AI agents as tools that augment human capabilities rather than attempting full job replacement. This means deeply understanding human workflows and identifying specific tasks AI can enhance to create more valuable and socially acceptable products.
Customer-Back AI Innovation: The call for "customer-back engineering" in AI underscores the importance of starting with user needs to drive breakthrough innovation. PMs must ensure their teams are deeply connected to customers to avoid building sophisticated AI solutions that don't solve real problems, leading to wasted investment.
Cost Efficiency for Mass Adoption: Advances like RateQuant, which drastically reduce LLM inference costs and memory footprints, are critical enablers for mainstream AI adoption and the viability of long-context applications. PMs should keep a pulse on infrastructure innovations that can unlock new product possibilities and improve the economic models of their AI offerings.
Mainstream AI Product Strategy: ChatGPT's broadening adoption across demographics signals that AI is moving beyond early adopters. PMs should adjust their product strategies to cater to a more diverse, mainstream audience, focusing on usability, practical applications, and clear value propositions for users who are not necessarily tech-savvy. For those looking to master building AI products for this evolving market, the PM Interview Prep Club's AI in Product Management course offers frameworks and tools to navigate these challenges.
Business Analyst at Landmark Leisure India · Yemalur, Bangalore
Product Owner at Availity · Bangalore (Remote, with onsite as needed)
Senior Product Manager at Pice · Bengaluru, Karnataka
Product Manager at StockGro · Bengaluru
Product Manager at ProsParity · Navi Mumbai, Maharashtra, India
Associate Product Manager at Lokal · Location Flexible
Senior Product Owner at Starrez · Hyderabad, Telangana
Product Manager-Digital Lending (Consumer Loans) at IDFC FIRST Bank · Mumbai, Maharashtra
Explore more openings on our PM job board or join the WhatsApp community for peer support and more.
If today's briefing was useful, forward it to a PM friend who'd appreciate it.
Thanks for reading PM Interview Prep Club! This post is public so feel free to share it.
Courses now live and launching soon:
AI in Product Management — Live now. Learn to lead AI product work: evaluate models, scope AI features, and drive AI-first roadmaps. Enrol today.
The 0-1 PM Playbook — Build products from scratch: discovery, MVP scoping, and go-to-market through real projects.
Growth PM, Technical PM, Design Thinking for PMs — Launching Soon. See all courses →
If you're in active prep: a Career Clarity session maps your path in one focused 1:1. The Pro Program gives you 3 months of mentored prep with real feedback. The Plus Plan lets you practice 500+ questions on your own schedule. Browse curated PM openings, try PM practice challenges, or join our free WhatsApp community for daily job alerts and peer support.
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.