TLDR 2026-03-31
Apple's AI strategy 🤖, Codex for Claude Code 🤖, Inside Sora's fall 🎥
Speak naturally. Send without fixing. (Sponsor)
Wispr Flow turns your voice into clean, sendable text in any app. Not rough transcription you have to fix. Text you actually send.
- 4x faster than typing. Emails, Slack threads, docs, AI prompts, brainstorms. Dictate anything, anywhere.
- 89% of messages sent with zero edits. Flow strips filler words, fixes grammar, and formats as you speak.
- Works in every app on every device. Mac, Windows, iPhone, and Android. One tap, start talking.
Millions of people use Flow daily, including teams at OpenAI, Vercel, and Clay.
Try Wispr Flow Free | Get Started
Apple Pivots Its AI Strategy to App Store, Search-Like Platform Approach (13 minute read)
Apple is recommitting to its core business model of selling hardware and services. It knows its homegrown AI technology lags behind competitors, so it is focusing on its core strengths. Historically, Apple's software has been about driving product sales rather than generating revenue in their own right. It plans to embed just enough AI in its OSes to keep users from defecting, while opening Siri and Apple Intelligence to third-party services. This approach leverages its hardware, makes its products more customizable, and keeps the company in control of its ecosystem.
The Sudden Fall of OpenAI's Most Hyped Product Since ChatGPT (14 minute read)
Sora was hyped as AI's next consumer-friendly frontier, and even Disney signed onto the vision. OpenAI then suddenly killed the project, with many Disney executives learning about the decision less than an hour before it was announced. OpenAI needed to free up more computing resources, and Sora was using far too much. The product wasn't profitable, and every user drew down a finite resource. CEO Sam Altman says the move was a difficult but necessary sacrifice toward the company's larger goals.
🚀
Science & Futuristic Technology
Vulnerability Research Is Cooked (16 minute read)
Coding agents will soon drastically alter both the practice and the economics of exploit development. A substantial amount of the high-impact vulnerability research will just be pointing an agent at a source three and telling it to find zero days. This will profoundly alter information security and the Internet itself.
Inside the stealthy startup that pitched brainless human clones (29 minute read)
A Californian startup called R3 has raised funds to create non-sentient monkey 'organ sacks' as an alternative to animal testing. The startup's founder reportedly secretly pitched a vision where the startup creates 'brainless clones' to serve as backups for human bodies. This could be in the form of a baby version of a person with enough brain structure to be alive in case they ever need a new kidney or liver, or it may even be possible to one day transplant a brain into a younger clone to gain a second lifespan. The company has denied that it ever made the pitch, saying that any allegations of intent or conspiracy to create human clones or humans with brain damage are categorically false.
💻
Programming, Design & Data Science
Kestra [26K GitHub stars]: The automation crisis ends here. (Sponsor)
Your stack is held together by legacy cron jobs, scripts, and fragile pipelines—and it's breaking. Kestra is an open-source, event-driven orchestrator to run your infrastructure, data, APIs, and AI workflows in one place. Define everything as code. Trigger in real time. Scale to millions. No glue code. No duct tape. Just reliable orchestration. Used by Apple, JPMorgan Chase, and Toyota.
Discover Kestra.Codex plugin for Claude Code (GitHub Repo)
This plugin enables Claude Code users to start using Codex from the workflow they already have. Users can use Codex from inside Claude Code for code reviews or delegate tasks to Codex. The plugin uses the local Codex CLI, so existing sign-in methods and configs still apply. It also uses the same repository checkout and machine-local environment.
Engineering for Agents That Never Sleep (2 minute read)
Most of the signals that trigger engineering work are already machine-readable. The current bottleneck is where a human reads that signal, interprets it, and then types a prompt telling the agent what the signal already said. Give agents comprehensive unit tests so they can verify their own output, good documentation to discover context, reproducible dev environments to run code, and rich system accounting. This scaffolding will enable agents that can close incidents around the clock.
Seeing like a spreadsheet (28 minute read)
AI will see much further than the spreadsheet ever could. It will allow businesses to accomplish incredible things, and make possible coordinations of labor and capital more ambitious than anything we can imagine today. However, as corporations become dominated by AI systems, the most human elements of organizations may be devalued or discarded entirely. Great organizations are great because of something irreducible about the collection and organization of particular people toward particular ends.
To Lure Top AI Talent, Startups Are Turning to Cold Hard Cash (6 minute read)
High-growth AI startups are flush with venture capital. Combined with a wildly competitive talent market, this has caused increasingly creative incentive structures and more cash-heavy offers. Startups are focusing on hiring the best as they need to stay lean and small. This is creating a split economy in the tech industry where the top 5% to 10% of candidates are getting all of the offers and the rest are struggling.
Make smarter application decisions with Azure Copilot Migration Agent (Sponsor)
Stop guessing what to modernize. This playbook from Microsoft's Azure Copilot Migration Agent team gives you strategies to decide what to re-platform, refactor, and what to leave as is.
Try it yourself–download a copy today.
🎁 Get 1 free month of Granola - the AI notepad you keep hearing about (Sponsor)
Granola transcribes directly from your device audio, so no bot ever joins the call. Jot down what matters, and Granola makes your notes awesome.
Claim your free month with code: TLDR1MO
Your ticket is a prompt (2 minute read)
Tickets are now handled by agents, and they only have the words on the ticket, and no other context.
Apple steps up crackdown on vibe coding apps, pulls ‘Anything' from the App Store (3 minute read)
Apple is not okay with apps that can change their own behavior by generating and running code outside of App Store review.
AI Agents Could Make Free Software Matter Again (30 minute read)
AI will create more software that is free to run, study, modify, and share.
Veblen & Jevon Walk Into a Data Center (3 minute read)
Jevon's Paradox states that the cheaper a product becomes, the more it is consumed, and Veblen goods are those whose demand increases with price - it is unknown which way the AI industry will go.
Meta testing paid Instagram service that lets users secretly view Stories (1 minute read)
Instagram Plus is being tested in select markets, including the Philippines, Japan, and Mexico, with the plans costing the equivalent of $2 USD per month.
Cost to Implement vs Cost to Verify (11 minute read)
The cost to implement is the time and expertise needed to produce code, and the cost to verify is the time and expertise needed to confirm the code is correct.
Get the most interesting stories in startups, tech, and programming delivered in a free daily email.
Join 1,600,000 readers for
one daily email