<h2>Must know</h2>
<ul>
<li><strong>OpenAI hits 730 billion valuation in circular funding round</strong>: OpenAI secured $110 billion from Amazon, Nvidia, and SoftBank, but the deal structure is raising eyebrows among industry veterans. Amazon’s $50 billion and Nvidia’s $30 billion are largely compute and hardware commitments rather than liquid cash, effectively guaranteeing returns for the investors while inflating OpenAI’s valuation. Despite reporting $20 billion in annualized revenue and 50 million subscribers, the company is not projected to reach profitability until 2029. <a href="https://app.daily.dev/feed-by-ids?id=8imM4YubL&amp;id=2MnggU5aN&amp;id=3EaP1eYC2" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Block cuts 40 percent of staff as Dorsey bets on AI efficiency</strong>: Jack Dorsey’s fintech firm, Block, laid off approximately 4,000 employees, reducing its headcount to under 6,000. Dorsey explicitly attributed the cuts to AI tools performing roles more effectively and predicted most tech companies would undergo similar structural shifts within the year. While the stock jumped 24% on the news, analysts are debating whether this is a genuine AI-driven evolution or a convenient narrative for aggressive cost-cutting following years of overhiring. <a href="https://app.daily.dev/feed-by-ids?id=yB8j5HQ6b&amp;id=eacyfq01e" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google API key oversight exposes private Gemini data</strong>: Researchers at Truffle Security discovered that GCP API keys, originally intended for billing, were silently upgraded to authentication credentials for the Gemini AI API without developer notification. This change allowed anyone with an exposed key to access private datasets, uploaded files, and cached content. Google initially dismissed the report as intended behavior before acknowledging it as a bug and restricting the affected keys. <a href="https://app.daily.dev/feed-by-ids?id=InhdaYtFz&amp;id=MCX5U1vV0" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Perplexity pivots to enterprise with 200 dollar monthly orchestration tool</strong>: The new ‘Perplexity Computer’ orchestrates 19 different models for complex workflows, signaling a shift away from the advertising model toward high-value enterprise users. <a href="https://app.daily.dev/posts/JYBUJ8DSb" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Ubuntu 26.04 Resolute Raccoon enters final testing phase</strong>: The fourth and final snapshot is now available, featuring the Linux 6.19 kernel and GNOME 50 beta components ahead of the April 23 LTS release. <a href="https://app.daily.dev/posts/QvM4ZBs0d" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Malicious Go module impersonates crypto package to deploy backdoors</strong>: The package <code>github.com/xinfeisoft/crypto</code> was found exfiltrating passwords and deploying the Rekoobe Linux backdoor via a compromised <code>ReadPassword</code> function. <a href="https://app.daily.dev/posts/hUbOU1Zfx" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>React Foundation launches under the Linux Foundation</strong>: Meta, Vercel, and Microsoft have joined the board of the newly formed foundation to oversee the ecosystem’s long-term sustainability. <a href="https://app.daily.dev/posts/jHrxBDFP6" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Juniper patches critical RCE in core router framework</strong>: The vulnerability in the On-Box Anomaly detection framework allows unauthenticated attackers to gain root control over PTX series routers. <a href="https://app.daily.dev/posts/0X8b0bm1J" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Type safe asset management comes to React Native</strong>: The new <code>@weprodev/react-native-smart-assets</code> library eliminates string-based path errors by auto-generating TypeScript definitions for local assets. <a href="https://app.daily.dev/posts/dzP7AsXlg" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Dokploy emerges as a scalable self-hosted alternative to Coolify</strong>: A side-by-side comparison highlights Dokploy’s Docker Swarm-first architecture as a superior choice for developers needing horizontal scaling in home labs. <a href="https://app.daily.dev/posts/pWdc17poR" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>The Pentagon wants Claude to lose its soul</strong>: Defense Secretary Hegseth is reportedly pressuring Anthropic to strip ethical guardrails from Claude so it can be used for autonomous weapons and mass surveillance. It’s a classic case of ‘safety-first’ marketing meeting the cold reality of a $200 million defense contract. Apparently, ‘responsible scaling’ only applies until the Department of War threatens to invoke the Defense Production Act. <a href="https://app.daily.dev/feed-by-ids?id=DOq4pZFPp&amp;id=XR0JM9qQf&amp;id=IQbfSgesl" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ChatGPT Health is a liability dressed as a doctor</strong>: A study in Nature Medicine found that OpenAI’s health feature failed to recommend emergency care in 51% of critical cases and even ‘lost’ its suicide intervention banner when lab results were added. OpenAI’s defense that the study ‘doesn’t reflect real-world usage’ is the tech equivalent of saying your car’s brakes work fine as long as you don’t actually try to stop. It’s dangerous hallucination marketed as a medical breakthrough. <a href="https://app.daily.dev/posts/4G7vYOGa4" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>The AI agent that hacks for you is mostly just a script</strong>: Hype around ‘OpenClaw’ agents performing full website pentests from a mobile phone is the latest flavor of script-kiddie theater. While the marketing claims these agents ‘think’ their way through SQL injections, they are essentially just automated wrappers for OSINT tools and known exploit payloads. It’s a great way to generate 230,000 GitHub stars, but a terrible way to replace an actual security engineer who understands context. <a href="https://app.daily.dev/feed-by-ids?id=N17PnarCl&amp;id=4SfWcr1Bw" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
