- Future//Proof
- Posts
- 😳 AI teams now ship 6.5× more work, and got back exactly zero hours
😳 AI teams now ship 6.5× more work, and got back exactly zero hours
💸 Stripe buys the AI cost layer for $7B+, Alipay turns merchants into agent-ready storefronts, and one link could have drained your Copilot inbox.
Avinash Mada
August 20, 2026
Welcome to the Future//Proof 🚀
👋 Hello , the AI Enthusiast.
In this week’s edition, we brought AI updates backed by high-quality research and data to give you deeper insights. You'll find the Top AI Breakthrough of the Week, a featured AI tool with a mini-tutorial, learning resources to help you master these tools, the top 3 AI news stories, and more.
Our goal is to help you improve your knowledge and stay ahead in the rapidly evolving AI landscape. You can submit your questions, queries, thoughts, opinions or anything regarding AI as a reply to this email and we'll feature and address them in our next newsletter.
🚀 Now Let’s dive in and explore the new AI Insights together!
⌛ Read Time: 6m:48s
How Jennifer Aniston’s LolaVie brand grew sales 40% with CTV ads
The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.
An in-depth look at a major AI development, its industry impact, how it could affect your career, and a bold future prediction.
AI now authors half the work, and teams are still working longer
Linear, the product-development platform used by tens of thousands of software teams, published "How teams build," its first data report, drawing on aggregated usage across roughly 199,000 paid users and 47,900 workspaces.
AI adoption more than doubled in every single function between January and June 2026. Product climbed from 12% to 34%, engineering from 12% to 30%, design from 6% to 22%, and even go-to-market, the function furthest from any codebase, went from 5% to 18%. CEOs at companies of 201+ people jumped from 9% to 36%, the single biggest leap in the report.
Output followed. Pull requests per workspace are up 111% against a June 2024 baseline. Teams that connected a coding agent went from 21 per week to 65. Teams that didn't went from 8 to 10, a 6.5× gap by Linear's own framing. AI now authors just under half of everything created in Linear, up from fewer than one item in a thousand two years ago. Non-engineers are shipping code too: product managers attaching pull requests rose from 3% to 10%, designers from 1% to 8%, founders from 11% to 23%.
Then comes the twist. Time spent creating, triaging, assigning and commenting went up in nearly every function. Planning stayed flat. AI chat and agent supervision appeared as an entirely new layer of work, and nothing shrank to make room for it. Linear's own conclusion is blunt: teams are working more, not less, with what it calls a Jevons-paradox quality to the whole thing.
Potential Impact
This is the cleanest public evidence yet that AI is a throughput technology, not a time-saving technology, at least as most organisations currently deploy it. Three consequences stand out:
The business case breaks at step three. Every AI pitch assumes: adopt AI, tasks get faster, spare capacity appears, costs fall. The tasks did get faster. The spare capacity never showed up, because work expanded to fill it.
Coordination is the new cost centre. More output means more review, more context-setting, more decisions about which AI output is good enough. That overhead is invisible on any dashboard.
Roles are collapsing. When founders, PMs and designers all ship code, and CEOs use AI more than their teams do, the line between deciding and executing dissolves.
For businesses, the question shifts from should we adopt AI to are we actually harvesting what it produces.
Implications for People/Careers
If you were promised your evenings back, this is your reality check. AI raises your ceiling, not your free hours. The professionals pulling ahead aren't doing the same job faster, they're doing a bigger job because execution stopped being the bottleneck.
Volume is no longer a differentiator. If everyone can produce more, producing more stops being impressive, and judgment, prioritisation and knowing what not to build become the scarce skills. Meanwhile the "non-engineer shipping code" trend means role boundaries are now yours to redraw: a marketer who builds a working tool, an ops lead who automates a process, a designer who ships a fix.
The career opening is real and unevenly taken. New LinkedIn data released this week shows women made up just over a quarter of new US hires into AI roles last year, versus roughly half of hires in non-AI occupations. These are the fastest-growing, best-paid roles in an otherwise sluggish market, and the barrier is usually a portfolio of applied work, not a degree.
Our Future//Take
The 2023 to 2025 story was can AI do the work? That answer is now clearly yes. The 2026 story is can you capture the value?, and this data suggests most organisations can't yet.
The next competitive divide won't be between companies that use AI and companies that don't. Adoption is already close to universal. It will be between companies that redesign the workflow and companies that bolt AI on top of the old one. The first group converts speed into margin. The second just gets faster at being busy.
The most valuable discipline for the next twelve months is deceptively simple: for every task you hand to AI, delete the manual step it replaced. If you can't delete it, you haven't automated it. You've duplicated it. Here’s your ₹25,000 AI Gift for FREE 🎁
Quick summaries of this week's top AI news, their relevance to your career, and our expert opinions.
Stripe confirmed on August 19 that it's acquiring OpenRouter, the gateway that lets developers reach hundreds of AI models through a single API. Bloomberg reported the deal at more than $7 billion, over five times the $1.3 billion valuation OpenRouter carried after its $113 million Series B in May. OpenRouter says it serves around 8 million users across 400+ models, with weekly token volume up roughly fivefold since the end of 2025. Stripe framed the logic around helping customers route requests intelligently and spend tokens efficiently.
Why It Matters to You
A payments company just paid billions for the layer that decides which model answers your request. That tells you where margin is moving. As AI shifts from pilot to production, inference stops being a rounding error and becomes a line item you manage like payroll or ad spend. If your business routes everything through one provider because that's what you wired up in 2024, you are very likely overpaying for tasks a cheaper model handles identically.
Our Take
Single-vendor AI is becoming a liability on three fronts at once: price, availability and negotiating leverage. Expect "AI FinOps" to become a real function inside mid-sized companies within a year, and expect model routing to become a standard feature of every automation platform rather than a specialist tool. The businesses that benefit are the ones that already know which of their tasks are simple and which are hard.
At its AI Ecosystem Partner Conference in Hangzhou this week, Alipay unveiled what it calls China's first full-stack agentic commerce platform. Merchants can convert their web pages, product catalogues and service workflows into agent-ready "Skills" and MCP tools, machine-readable capabilities an AI agent can actually call. Those merchants then plug into Ah Bao, Alipay's consumer agent launched in June, through an interoperability protocol called AHA. KFC, Luckin Coffee and Mixue are already integrated, alongside five smartphone brands representing over 70% of the market and 16 automakers. To seed adoption Alipay is handing out 100 million free tokens per participant plus transaction subsidies. Alibaba shares rose as much as 5% in Hong Kong. Ant Group's CEO expects agentic commerce to grow sharply over the next 6 to 12 months.
Why It Matters to You
For twenty years, being findable meant SEO, optimising for humans reading search results. Agentic commerce changes who the reader is. If an AI agent does the shopping, booking and paying, your business needs to be readable by software, not just persuasive to people. A beautiful landing page an agent can't parse is effectively invisible.
Our Take
This is China-first, but the pattern is global. Stripe, OpenAI and PayPal have all been laying agent-to-merchant payment rails. The practical move for any business owner is boring but valuable: make sure your pricing, availability, service descriptions and booking flows exist as structured, machine-readable data, not just as images and marketing copy. This is the 2026 equivalent of getting your Google Business Profile right in 2012, and the businesses that do it early will be the ones agents can actually transact with.
On August 18, Microsoft shipped a fix for CVE-2026-24301, nicknamed CoSnitch by the Varonis Threat Labs researchers who found it. The flaw chained an undocumented URL parameter, Copilot's built-in ability to fetch URLs, and persistent memory poisoning. The result: a single malicious link could auto-execute prompts and exfiltrate connected Gmail, Drive and Calendar data, with no click-through and no confirmation. CSO Online reports the fix landed more than eight months after Varonis first disclosed it.
Why It Matters to You
Every AI assistant you connect to your inbox, drive or calendar is a new door into your business data. Traditional security training taught people not to open suspicious attachments. This class of attack works by feeding instructions to your AI, which then acts with your permissions. Your staff are no longer the attack surface. Your integrations are.
Our Take
Treat AI connectors like employee access, because functionally that's what they are. Audit what each AI tool can read. Use least-privilege connections rather than blanket account access. Turn off persistent memory for assistants that touch sensitive data. As agents move from answering questions to taking actions, prompt injection stops being a research curiosity and becomes an operational risk with a cost attached. If a vendor can't tell you exactly what an assistant is permitted to reach, that silence is your answer. AI Mastery for FREE (Sign up Now).
Discover a comprehensive guide to an AI tool, exploring its features, practical use cases, and learning resources to help you master it.
Given that Stripe just paid billions for it, this week's tool picks itself.
OpenRouter is a single API gateway that gives you access to hundreds of AI models from 60+ providers, including OpenAI, Anthropic, Google, Meta, Mistral, DeepSeek and a long tail of open-weight labs, through one key, one endpoint and one prepaid credit balance. It's the plumbing layer between your application and the model market. Unlike a workflow tool such as n8n or Zapier, OpenRouter doesn't automate anything. It decides which model answers each request, and at what price.
The reason it matters for business owners and not just developers: it turns model choice from an architecture decision into a per-request setting. You stop being locked into whichever provider you happened to sign up with first.
⭐ Top Features
One key, one bill. Replaces separate accounts, billing dashboards and API keys for every provider with a single balance and unified usage analytics.
OpenAI-compatible endpoint. If your code, automation or n8n workflow already calls OpenAI, you usually only change the base URL. Migration cost is close to zero.
Automatic fallback routing. List models in priority order. If your first choice is rate-limited or down, the request falls through to the next instead of failing. For anything customer-facing, that's real uptime insurance.
Price-first provider routing. By default it sends requests to the cheapest reliable provider serving that model, and you can override to optimise for speed instead.
Hard price caps. Set a maximum price per request and it simply won't execute if no provider meets your threshold. A genuine guardrail against runaway agent spend.
Free model variants. Rate-limited, but enough to prototype a workflow before you spend anything.
BYOK. Keep your existing direct provider contracts and use OpenRouter purely as the routing and failover layer.
Resources for Learning
Official Quickstart: openrouter.ai/docs/quickstart for creating your key and sending your first request.
Model Catalogue: openrouter.ai/models for live per-token pricing across every provider, so you can price a task before you commit to it.
A curated list of noteworthy AI tools and their key details to help you stay ahead in your field.
Cursor's August 19 release turned its cloud agents into always-on workers. A new Subscriptions system lets agents monitor pull requests, watch Slack threads and run scheduled tasks, waking when something happens rather than waiting to be summoned. It added subagents on isolated VMs with their own project copies, and a goal command for long-lived objectives like "fix all flaky tests and keep CI green." Its Router also splits each request between quality and cost modes, which is the three-tier logic above applied automatically. The standout is persistence: this is the shift from agent-as-tool to agent-as-teammate, which means you now need scope, guardrails and a review cadence rather than just a prompt.
A no-code, visual AI workflow and agent builder aimed squarely at ops, marketing, sales and support teams. You drag AI reasoning nodes onto a canvas alongside app integrations, add human-in-the-loop approval steps where judgment matters, and run flows on a trigger or a schedule. It supports MCP, which matters given where agentic commerce is heading. The differentiator is inspectability: you can see exactly what each step is doing and where it went wrong, which is the difference between an automation you trust and one you babysit.
Think of it as a spreadsheet with research superpowers for go-to-market teams. It pulls from multiple data providers to enrich messy lead lists, and its Claygent feature reads public web pages to answer specific questions you define, such as "do they have a pricing page" or "what market do they serve," then drops the answer into a column. The differentiator is research as a column: instead of a filled sheet, you get a list you can actually write personalised outreach from. Ask narrow questions and expect short answers for the cleanest results.
The fastest route from a rough idea to something you'd actually put in front of a client. Describe a deck, paste an outline, or upload a document, and Gamma returns a designed, web-native result in under a minute. It covers presentations, documents, websites and social graphics from the same platform, and exports to PPTX, PDF and Google Slides. The standout is the Gamma Agent, a conversational layer that handles deck-wide edits from a single instruction, so "make this more corporate" or "add a competitor comparison" restyles the whole thing instead of you clicking through every card. Watch the credit system on lower tiers during revision-heavy sessions, and check PPTX exports before you send anything mission-critical.
A quick poll to help you recollect and engage with key points from the newsletter.
Linear's data showed AI adoption doubling and output rising 111%. What happened to the total time teams spent on product development? |
Share your feedback on today's edition to help us improve and better meet your needs.
How was Today’s Edition? |
Share our Newsletter ⏩
Enjoying our insights on the latest AI breakthroughs? Don’t keep it to yourself! Share this newsletter with friends and colleagues who are passionate about technology and AI innovation.
If you haven’t subscribed yet, make sure to subscribe here to stay updated with cutting-edge AI news, tools, and tutorials delivered straight to your inbox!
Ask Us Anything AI ❓
Got questions? We've got answers!
Submit your questions, queries, thoughts, opinions or anything regarding AI and we'll feature and address them in our next newsletter. Your curiosity drives our content!
👇 Reply to this email with your questions, and we'll answer them in our next edition!👇