<h2>Must know</h2>
<ul>
<li><strong>OpenAI goes all-in on ads, Musk sues for $134 billion</strong>: OpenAI is introducing targeted ads to its free and Go ($8/month) ChatGPT tiers in the US, a move CEO Sam Altman previously called a ‘last resort.’ This shift aims to diversify revenue as the company faces massive losses, reportedly over $11.5 billion in Q3 2025 alone, with some sources projecting a $45 billion cash burn by 2028. Meanwhile, Elon Musk is suing OpenAI and Microsoft for up to $134 billion, alleging breach of OpenAI’s original nonprofit mission and claiming entitlement to compensation for his early contributions, as reported by Hacker News and business press. The lawsuit is headed to a jury trial in late April. <a href="https://app.daily.dev/feed-by-ids?id=DQh9IAXNl&amp;id=ep7p1hnQy&amp;id=pwpD2QRjm" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI coding tools are making developers slower and buggier</strong>: Despite the hype, AI coding assistants are reportedly making developers 19% slower and doubling bugs, according to The Serious CTO. AI-generated code contains 1.7 times more security issues and increases pull request review time by 91%. The core problem is that AI optimizes for typing speed over system design, leading to technical debt and developers merging code they don’t fully understand. This exposes a critical need for human judgment in architecture and fundamental understanding that AI currently cannot replicate. <a href="https://app.daily.dev/feed-by-ids?id=5khP9Wt37&amp;id=jSgoQLSoF&amp;id=3sM5oeueQ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google Gemini calendar flaw exposed enterprise data</strong>: Google Gemini had a security vulnerability that allowed attackers to extract private calendar data through prompt injection. Malicious natural language prompts embedded in Google Calendar invitations could bypass privacy controls when users asked Gemini about their schedules. This flaw, discovered by security researchers, highlights emerging security risks unique to LLM-powered applications, especially in enterprise environments where LLMs connect to sensitive systems, as reported by Security Boulevard. Google has since patched the vulnerability. <a href="https://app.daily.dev/feed-by-ids?id=dHjNBMkFO&amp;id=j6bBmLvCW" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>AI is shifting software engineering roles</strong>: AI coding agents are accelerating implementation, shifting the bottleneck from coding to problem definition, creating a ‘Product Engineer’ role that blends technical skills with deep customer understanding and strategic thinking, according to Atlassian and InfoWorld. <a href="https://app.daily.dev/feed-by-ids?id=ruSCTD0Fh&amp;id=sifgyKGyd&amp;id=H430dwVHX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic launches Claude Co-work for desktop automation</strong>: Anthropic launched Claude Co-work, expanding Claude Code’s capabilities to general desktop automation via a chat interface, enabling non-technical users to automate tasks like file management and terminal commands on Mac, though it faces an unpatched prompt injection vulnerability that allows file exfiltration. <a href="https://app.daily.dev/feed-by-ids?id=3e5iSak1o&amp;id=s11rN52dz&amp;id=jaKSkIh92&amp;id=rAHVvRPEj" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Small language models gain enterprise preference</strong>: Enterprises are increasingly favoring Small Language Models (SLMs) over Large Language Models (LLMs) due to lower costs, reduced latency, and better fit for closed, well-defined workflows like invoice parsing and contract classification, with models like Phi-3 and Mistral 7B matching or surpassing LLMs on constrained tasks. <a href="https://app.daily.dev/feed-by-ids?id=HX43N3MkE&amp;id=J1BKrq2u6" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GitHub Copilot gets repository-specific memory</strong>: GitHub Copilot introduced a ‘memories’ feature in Visual Studio, automatically learning and documenting coding preferences and project guidelines to maintain code consistency, speed up onboarding, and automate enforcement of team standards, with a public preview available across all paid Copilot plans. <a href="https://app.daily.dev/posts/nCmOnsQS3" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft open-sources XAML Studio</strong>: Microsoft has open-sourced XAML Studio under the .NET Foundation, allowing rapid prototyping and fast feedback loops for XAML-based UI development, with the latest version (2.0) featuring a UI redesign based on Fluent principles and enhanced integration with WinUI 3. <a href="https://app.daily.dev/posts/hUJrXK08w" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cloudflare acquires Astro web framework</strong>: Cloudflare acquired The Astro Technology Company, maker of the Astro web framework, which will remain open source, MIT-licensed, and platform-agnostic, with the entire Astro team joining Cloudflare to continue development. <a href="https://app.daily.dev/feed-by-ids?id=a23rECuU6&amp;id=6Yvvs8a9F&amp;id=yRAOJZmiR" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New Python SDK for application-layer security</strong>: Arcjet released a Python SDK (beta) for application-layer security, embedding rate limiting, bot detection, and spam prevention directly into Python codebases for FastAPI and Flask, with code running locally via WebAssembly for efficiency. <a href="https://app.daily.dev/posts/Kk7M89j0h" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI video startup lands $1.3B valuation with 15M users</strong>: Higgsfield, an AI video generation startup, just raised $80 million at a $1.3 billion valuation, claiming 15 million users and a $200 million annual revenue run rate just nine months after launch. Apparently, we’re all just churning out AI-generated videos now, and it’s worth billions. Who needs actual content when you can have ‘controversial use cases’ and a massive valuation, right? <a href="https://app.daily.dev/posts/57X1LVeos" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Razer CEO defends AI investments in gaming</strong>: Razer’s CEO, Min-Liang Tan, is defending their $600 million AI investment, including ‘Project Ava’ (an AI companion using Grok), by claiming gamers only reject ‘generative AI slop’ but value AI for quality assurance. So, apparently, the problem isn’t AI, it’s just <em>bad</em> AI. He also acknowledges ‘mental health concerns’ and ‘unresolved trust and safety issues’ with Grok, but hey, $600 million says they’ll figure it out, eventually. <a href="https://app.daily.dev/posts/bokhtiGjg" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI data centers face skilled worker shortage</strong>: The AI data center boom is creating an ‘unprecedented demand’ for skilled workers, with the U.S. needing nearly 400,000 more construction workers by 2033. So, while AI is supposedly making everyone redundant, we’re now scrambling to find enough humans to build the infrastructure for it. It’s almost like the robots can’t lay their own fiber optic cables yet. Who knew? <a href="https://app.daily.dev/posts/JE89HgTw1" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
