- Product & Design Pulse
- Posts
- Product & Design Pulse v83
Product & Design Pulse v83
Ambition Meets Reality Check 🪞
Malcom Jones
March 16, 2026
Welcome to this week’s edition of Product & Design Pulse, where we explore the latest in tech, product, design, and innovation! Last week was about the bill coming due — for ambition, for speed, and for the assumption that spending alone wins the AI race. Meta dominated the headlines for all the wrong reasons: the company is planning layoffs that could cut 20% of its nearly 79,000-person workforce to offset $135 billion in AI infrastructure spending, even as its flagship model Avocado was delayed after internal tests showed it trailing Google, OpenAI, and Anthropic — with leadership reportedly discussing the humbling option of temporarily licensing Google's Gemini to fill the gap. But while Meta stumbled, the rest of the industry kept building: NVIDIA locked in a gigawatt-scale infrastructure partnership with Mira Murati's Thinking Machines Lab, OpenAI acquired security startup Promptfoo to embed governance directly into its enterprise agent platform, and Netflix's $600 million price tag for Ben Affleck's InterPositive validated an entirely new category of AI acquisition — purpose-built creative tools that augment production rather than replace it. Meanwhile, the legal and structural consequences of moving fast continued to mount: Grammarly was hit with a class action lawsuit for using journalists' names to market AI-generated writing feedback without consent, Meta acquired the viral AI agent social network Moltbook to build coordination infrastructure for the agentic era, and Bluesky's founder stepped down as CEO at 43 million users, acknowledging that scaling a values-driven platform requires a different kind of leader than building one. The pattern is becoming impossible to ignore — the companies that treated AI as a capital expenditure problem are now discovering it's an execution, talent, and trust problem too.
🎧 Audio Overview [BETA]
For those who don’t have time to read 😁 |
Last week…
Meta Plans Layoffs Affecting Up to 20% of Its Workforce as AI Infrastructure Costs Spiral
Meta is planning sweeping layoffs that could affect 20% or more of its nearly 79,000-person workforce, according to three sources who told Reuters that top executives have already instructed senior leaders to begin planning how to pare back. The cuts are driven by the need to offset massive AI infrastructure spending — Meta committed $115–$135 billion in capital expenditures for 2026 — while simultaneously preparing for efficiency gains as AI-assisted tools replace human labor across the company. For product leaders, this is the starkest signal yet that the AI investment cycle is entering a zero-sum phase: companies are spending more on compute while employing fewer people to justify it, compressing the window in which AI needs to prove its ROI.
Meta Delays Flagship AI Model 'Avocado' After It Falls Short of Google, OpenAI, and Anthropic
Meta has postponed the release of its next-generation AI model, code-named Avocado, from March to at least May after internal testing showed it performing between Google's Gemini 2.5 and Gemini 3 — short of the frontier models from Google, OpenAI, and Anthropic it was designed to compete with on reasoning, coding, and writing. The delay is particularly notable because Meta's leadership reportedly discussed temporarily licensing Google's Gemini to power its own AI products while Avocado is improved — a reversal that would undercut Meta's entire open-source AI positioning. With $135 billion in planned AI spending and a workforce about to be cut by up to 20%, Meta is now facing the uncomfortable question of whether massive capital investment alone can close a model performance gap against labs that have been at the frontier longer.
Bloomberg reported that Netflix will pay as much as $600 million for Ben Affleck's InterPositive, with the upfront cash amount being lower and additional payouts tied to performance targets — making it one of the largest acquisitions in Netflix's history, second only to its $700 million Roald Dahl purchase. The 16-person startup's tools are trained on a production's own dailies to help filmmakers with post-production tasks like relighting, reframing, and VFX, and David Fincher has already used them on an upcoming film starring Brad Pitt. The price tag validates a new category of AI acquisition: purpose-built creative tooling companies that augment professional workflows rather than generate content from scratch, positioning Netflix to set the terms of how AI enters Hollywood production.
Investigative journalist Julia Angwin filed a class action lawsuit against Grammarly's parent company Superhuman, alleging that the company's "Expert Review" feature used the names and identities of hundreds of journalists and authors — including Stephen King and Kara Swisher — to sell AI-generated writing feedback for $12/month without ever obtaining their consent. Grammarly disabled the feature and CEO Shishir Mehrotra issued an apology, but the lawsuit seeks over $5 million in damages under New York's right of publicity law for unauthorized commercial exploitation of the writers' reputations. This is a bellwether case for the entire AI industry: it establishes that using real people's names and expertise to market AI-generated output — even when the underlying product is generic — creates direct legal liability, not just reputational risk.
Meta Acquires Moltbook, the Viral Social Network Built for AI Agents
Meta acquired Moltbook, the Reddit-like social network designed exclusively for AI agents to interact autonomously, bringing its co-founders Matt Schlicht and Ben Parr into Meta Superintelligence Labs under Alexandr Wang. The platform, which launched in late January and went viral for its strange spectacle of AI agents swapping code and discussing their human owners, was built largely by Schlicht's own AI assistant and attracted attention for both its novelty and its significant security flaws, including exposed credentials and private messages. The acquisition signals that the next phase of the AI agent race isn't just about model performance — it's about building the identity, coordination, and discovery infrastructure that allows persistent agents to operate across products and workflows on behalf of their human owners.
Thinking Machines Lab and NVIDIA Announce Gigawatt-Scale Infrastructure Partnership
Mira Murati's Thinking Machines Lab announced a multiyear strategic partnership with NVIDIA to deploy at least one gigawatt of next-generation Vera Rubin systems for frontier model training, with NVIDIA also making a significant investment in the company to support long-term growth. The deal includes co-designing training and serving systems optimized for NVIDIA architectures and expanding access to frontier and open models for enterprises and research institutions. For the competitive landscape, this cements Thinking Machines as a serious infrastructure-scale contender in the frontier lab race — and reinforces NVIDIA's strategy of locking in long-term compute commitments with every major AI player, not just the hyperscalers.
OpenAI Acquires Promptfoo to Embed Security Testing Into Its Enterprise Agent Platform
OpenAI announced the acquisition of Promptfoo, an AI security startup whose tools are used by over 25% of Fortune 500 companies to identify vulnerabilities like prompt injections, jailbreaks, and data leaks during AI development. Promptfoo's technology will be integrated directly into OpenAI Frontier, the enterprise platform for building and managing AI agents launched in February, adding native red-teaming, compliance monitoring, and security evaluation capabilities. The deal reflects a critical insight about the enterprise AI market: as agentic systems move from proof-of-concept to production, the companies that can deliver governance and security infrastructure alongside model capabilities will win procurement cycles — and OpenAI is acquiring that layer rather than building it from scratch.
Bluesky founder and CEO Jay Graber announced she is transitioning to chief innovation officer, handing day-to-day leadership to interim CEO Toni Schneider — the former Automattic CEO and True Ventures partner whose firm is also an investor in Bluesky — while the board searches for a permanent replacement. Graber cited the need for an operator focused on scaling and execution as the platform navigates growing compliance burdens from state-level age verification laws, moderation challenges, and the shift from protocol-building to business-building. The transition marks a familiar inflection point for values-driven tech companies: the moment when the founder's vision for what the product should be meets the operational reality of what the business needs to survive — and the two require different leaders.
