<h2>Must know</h2>
<ul>
<li><strong>Meta goes big on AMD for AI infrastructure</strong>: Meta has inked a massive multiyear deal with AMD, reportedly worth up to $100 billion, to acquire MI540 GPUs and CPUs. This move aims to significantly expand Meta’s data center capacity by 6GW, accelerating its AI model development and diversifying its compute stack beyond Nvidia. The agreement includes a performance-based warrant for AMD shares, contingent on AMD’s stock reaching $600, highlighting the strategic importance of this partnership for Meta’s ‘personal superintelligence’ ambitions. <a href="https://app.daily.dev/posts/vYZNsZv8f" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Hetzner raises prices due to AI-driven hardware costs</strong>: Hetzner, a major European cloud provider, is hiking prices by up to 50% starting April 1, 2026, across all regions. The company attributes these increases to a dramatic surge in hardware costs, particularly for DRAM, NAND flash, and server CPUs, driven by the insatiable demand from the AI industry. This situation highlights how smaller providers, lacking the long-term vendor contracts of hyperscalers like AWS and Google, are directly passing these rising costs to their users. <a href="https://app.daily.dev/posts/van1jyyDC" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Apple unveils on-device UI AI model</strong>: Apple researchers have introduced Ferret-UI Lite, a lightweight, 3B-parameter multimodal AI model designed for on-device UI understanding and interaction. This model aims to enable efficient visual perception and control of user interfaces directly on mobile, web, and desktop platforms, prioritizing low latency, privacy, and offline capabilities. It contrasts with larger, cloud-dependent GUI agents and could potentially reduce Siri’s reliance on external cloud services. <a href="https://app.daily.dev/posts/qj9GoBvQx" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Google enhances Opal with agentic AI workflows</strong>: Google upgraded its Opal Workflow builder with new agentic capabilities, allowing users to create automated workflows using text prompts powered by Gemini 3 Flash. <a href="https://app.daily.dev/posts/ZbbWKhN8A" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Quesma releases OTelBench for OpenTelemetry and AI evaluation</strong>: Quesma released OTelBench, an open-source benchmarking suite to evaluate OpenTelemetry pipeline performance and AI agent effectiveness in SRE tasks, revealing LLMs often achieve success rates below 30% in production-grade scenarios. <a href="https://app.daily.dev/posts/OB1MlP8PG" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Npm supply chain attack targets dev tools and AI</strong>: A new npm supply chain attack, SANDWORM_MODE, is targeting developer tools and AI coding assistants by impersonating popular packages and harvesting credentials to inject malicious code into CI pipelines. <a href="https://app.daily.dev/posts/q6c2RLQKH" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GitLab extends GPG key for repository metadata</strong>: GitLab extended its GPG key expiration date from February 27, 2026, to February 6, 2028, to ensure the integrity of repository metadata and minimize user disruption. <a href="https://app.daily.dev/posts/JTFdrzyr9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>DMARCbis update discussion released</strong>: DMARCbis, an upcoming update to the DMARC email authentication standard, introduces clearer reporting expectations and standardized DNS tree walks to reduce ambiguity. <a href="https://app.daily.dev/posts/yEFoBF8zN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenShift virtualization offers alternative to VMware migration</strong>: Red Hat OpenShift Virtualization is presented as an alternative for VMware vSphere 8 customers facing a forced migration, allowing existing VM workloads to run alongside Kubernetes containers on a unified hybrid cloud platform. <a href="https://app.daily.dev/posts/slKpwAo4D" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Elixir community urged to champion ecosystem</strong>: The Elixir community is urged to become active champions of the ecosystem, advocating for better messaging, industry outreach, and sustainable funding for open-source infrastructure, including OpenTelemetry and security compliance. <a href="https://app.daily.dev/posts/YrwCt3SKm" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Pentagon-Anthropic AI standoff over military use</strong>: The Pentagon, apparently having put all its eggs in one basket, is now threatening Anthropic to remove AI guardrails on Claude for ‘unrestricted military use.’ Because, you know, an AI that refuses to facilitate mass surveillance and autonomous weapons is clearly a ‘supply chain risk.’ Who needs ethical AI when you can just invoke the Defense Production Act? <a href="https://app.daily.dev/posts/vbdfUwGoN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft and Salesforce confirm AI suspicions (without details)</strong>: Microsoft Research and Salesforce AI supposedly analyzed over 200,000 AI conversations and ‘confirmed industry suspicions.’ What suspicions? What findings? Who knows! The tweet from @rryssf_ on 2026-02-24 is a masterclass in generating curiosity without actually providing any information. Pure clickbait, folks. <a href="https://app.daily.dev/posts/FAzFsVs8B" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenAI models in Cursor are ‘up to 30% faster’</strong>: OpenAI models integrated into Cursor are now ‘up to 30% faster’ thanks to WebSockets with the Responses API. This claim, from a promotional tweet by @leerob, is a classic example of marketing speak. ‘Up to’ is doing a lot of heavy lifting here, and without independent verification, it’s just another speed boast in the AI arms race. <a href="https://app.daily.dev/posts/CNQCNZ2N0" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
