SpaceX drone swarms 🚁, Apple video podcasts 📱, AI isn't a bubble 🤖

SpaceX and xAI are competing in a secretive new Pentagon contract to produce voice-controlled autonomous drone swarming technology ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌  ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ 

Sign Up |Advertise|View Online TLDR Together With TLDR 2026-02-17 7 Security Best Practices for MCP (Sponsor) Learn what security teams are doing to protect MCP without slowing innovation. As MCP (Model Context Protocol) becomes the standard for connecting LLMs to tools and data, security teams are moving fast to keep these new services safe.  The MCP Security Best Practices Cheat Sheet outlines seven proven steps teams can put in place right away, including: How to lock down MCP servers and supply chains Enforcing least-privilege access for tokens and tools Adding human-in-the-loop safeguards for critical actions If you're starting to see MCP show up in your environment, this is a great place to start. Get the Cheat Sheet 📱 Big Tech & Startups SpaceX to Compete in Pentagon Contest for Autonomous Drone Tech (5 minute read) SpaceX and its wholly owned subsidiary, xAI, are competing in a secretive new Pentagon contract to produce voice-controlled autonomous drone swarming technology. While Elon Musk has been enthusiastic about advancing AI, he has previously argued for a ban on autonomous weapons that can select and engage their own targets and operate beyond meaningful human control. OpenAI is also working on the effort by supporting a successful submission from Applied Intuition. The company's contribution to the project will be limited solely to the conversion of voice and other instructions from battlefield commanders into digital instructions. Its technology won't be used for the operation of the drone swarm, weapons integration, or target authority. Apple takes on YouTube and Spotify with new video podcasting push (4 minute read) Apple will be introducing an integrated video podcast experience to Apple Podcasts in the spring. The update will bring Apple Podcasts more in line with its competitors, which have increasingly leaned into video podcasting. There will be a picture-in-picture mode, and users will be able to download video episodes for offline viewing. The update will introduce support for HTTP Live Streaming, which features dynamic video ad insertion. 🚀 Science & Futuristic Technology BCIs in 2026: Still Janky, Still Dangerous, Still Overhyped (17 minute read) The fundamental problems that plagued Brain-Computer Interfaces (BCIs) a decade ago are still here. While there is genuine, impactful progress in BCIs, this remains largely confined to severe medical conditions. These are highly specialized, medically critical applications, and there are severe engineering challenges that must be overcome before they can be extrapolated into consumer products. This article takes a look at the state of the BCI industry to provide a realistic look at where the technology currently stands and the path forward. Rare, dangerous side effects of some COVID-19 vaccines explained (6 minute read) Vaccine-induced immune thrombocytopenia and thrombosis (VITT) occurred in roughly one out of 200,000 people who received adenovirus-based COVID-19 vaccines. The condition occurred due to PF4 antibodies binding to rogue antibodies, setting off a dangerous cascade (PF4 is a protein involved in blood clotting). Researchers hypothesize that patients who developed VITT had previously been infected with an adenovirus, which primed their B cells to recognize an adenovirus protein in the vaccine, which kicked those cells into action, starting a mutation-generating process that produced a variant with a negatively charged amino acid, which ultimately led to VITT. In most parts of the world, between 40% and 60% of the population has the genetic background that makes people more prone to VITT, but in East Asia, the prevalence is only 20%. 💻 Programming, Design & Data Science Peak traffic shouldn't mean peak anxiety (Sponsor) Traffic skyrocketing? If you're on Microsoft Azure, there's no reason to panic. When user demand spikes, Azure adjusts capacity in the background. Performance stays steady with no manual intervention, no late-night firefighting, and no scrambling to explain why the site went down during the big launch. Build without limits with Azure Anthropic tries to hide Claude's AI actions. Devs hate it (3 minute read) Anthropic has changed Claude Code's progress output to hide the names of files the tool was reading, writing, or editing. Developers have pushed back, saying they need to see which files were accessed. The full details can still be accessed with a keyboard shortcut. Developers who want more details can also enable verbose mode. --dangerously-skip-reading-code (3 minute read) Large language models generate code much faster than humans can read it. We can't expect humans to effectively review, understand, and approve every diff anymore. Humans should focus their efforts on specifications and tests while agents code. Rework is almost free, so there should be less effort put into preventing incorrect work from happening. 🎁 Miscellaneous Why I don't think AI is a bubble (14 minute read) People mostly say that AI performance will plateau because they believe that performance increases have been driven by brute-force scale, and the trend can't last. However, there are also new innovations like reasoning and reinforcement learning driving improvements, so there's no real reason to believe progress will halt. The data centers being built now will likely be used productively. If the technology does actually work, it's reasonable to believe the industry will sort out any 'bubble'. I guess I kinda get why people hate AI (20 minute read) AI will likely become extremely helpful in the future, but right now, the everyday experience with the technology is awful. It may be helpful for programmers, but normal people tend to have more negative experiences, and none of the people behind the technology seem to care. AI continues to lower the barrier to entry for annoying, low-quality things. This is resulting in things becoming generally worse and is making the internet and social media and such increasingly unusable. ⚡ Quick Links Anam introduces cara-3: the world's most advanced interactive avatar model (Sponsor) Outperforms competitors up to 60%. Add a face to your product via API and improve your product's engagement by 44%. The Cost of Staying (6 minute read) The shift from execution to judgment is making the cost of staying and the upside of moving look completely different. When everything goes to 0 (9 minute read) Trust is a scarce resource and a moat. GeLLMan amnesia (5 minute read) When people talk about the jobs that AI will soon automate, they're always talking about somebody else's job. 10 Years Building Vertical Software: My Perspective on the Selloff (17 minute read) The market is finally distinguishing between companies that own something genuinely scarce that is safe from LLM agents. A Defector Explains the Remote-Work Scam Helping North Korea Pay for Nukes (8 minute read) North Korean IT workers, who were generally sent back to Pyongyang every two or three years for roughly a month of 're-education', would hand over 90% of their earnings back to the regime. Love TLDR? Tell your friends and get rewards! Share your referral link below with friends to get free TLDR swag! https://refer.tldr.tech/7c97ce4b/ Track your referrals here. Want to advertise in TLDR? 📰 If your company is interested in reaching an audience of tech executives, decision-makers and engineers, you may want to advertise with us. Want to work at TLDR? 💼 Apply here, create your own role or send a friend's resume to jobs@tldr.tech and get $1k if we hire them! TLDR is one of Inc.'s Best Bootstrapped businesses of 2025. If you have any comments or feedback, just respond to this email! Thanks for reading, Dan Ni & Stephen Flanders Manage your subscriptions to our other newsletters on tech, startups, and programming. Or if TLDR isn't for you, please unsubscribe.