Blog — Page 39
Engineering notes, case studies, and lessons from the field.
Microsoft's 'Rule of 70' Opens a Voluntary Exit for 8,750 U.S. Employees — Full Package Revealed May 7
Microsoft launched its first voluntary retirement program in 51 years, targeting workers whose age plus tenure totals 70 or more. The financial terms land on May 7, one week after Q3 earnings.
Intel's Xeon 7 'Diamond Rapids' Slips to Mid-2027 — AMD's EPYC Venice Gets an Uncontested Year in Data Centers
Intel has pushed its Xeon 7 'Diamond Rapids' server CPUs from 2026 to mid-2027, according to leaks corroborated across four publications on April 25. The platform — up to 512 cores, 16-channel DDR6, built on Intel 18A — now arrives a year late, handing AMD EPYC Venice a clear window in the data center market.
Shield AI Closes $2 Billion at a $12.7 Billion Valuation — Its Autonomous Pilot Is Already Flying for the US Air Force
Shield AI raised $2 billion across a Series G led by Advent International and preferred equity from Blackstone, then acquired simulation firm Aechelon Technology to own the full autonomous military AI stack. Its Hivemind software pilot has been selected for the US Air Force's Collaborative Combat Aircraft program.
CISA Adds Samsung MagicINFO RCE and SimpleHelp Flaws to KEV — All Four Under Active Exploitation, Patch Deadline May 8
CISA added four vulnerabilities to its Known Exploited Vulnerabilities catalog on April 25, setting a May 8 federal patch deadline. Samsung MagicINFO CVE-2024-7399 allows unauthenticated RCE; two SimpleHelp flaws chain into full server takeover; a D-Link DIR-823X command injection has no patch coming.
DeepSeek V4-Pro: 1.6 Trillion Parameters, 1M Context Window, and Benchmark Parity With Frontier Models at One-Tenth the Price
DeepSeek released preview access to V4-Pro, a 1.6-trillion-parameter MoE model with a genuine one-million-token context window that matches GPT-5 and Gemini 3.0 Pro on key benchmarks. At $0.145 per million input tokens, it reshapes the cost calculus for every enterprise LLM workload.
Rust Is Coming to CPython — Python 3.16 Target Set, PEP Draft Due June 2026
The CPython Rust integration team confirmed in its April 2026 update that CI builds are now passing on all platforms. A PEP draft lands in June–July 2026, with Python 3.16 beta in May 2027 and full release in October 2027. No Rust compiler required to build or run Python.
Pony.ai Debuts World's First L4 Autonomous Light Truck at Auto China 2026 — 40% Cheaper Than Human Drivers
Pony.ai unveiled a fully automotive-grade Level-4 electric light truck co-developed with CATL at Auto China 2026, alongside a Gen-7 robotaxi priced under $31,700 total. The company is deploying 1,000 Toyota bZ4X-based robotaxis across Chinese tier-1 cities this year.
CVE-2026-33626: LMDeploy SSRF Exploited 13 Hours After Disclosure — Upgrade to 0.12.3 Now
A high-severity SSRF vulnerability in LMDeploy's vision-language module was exploited in the wild just 12.5 hours after the GitHub advisory dropped. CVSS 7.5, affects all versions below 0.12.3. AI inference infrastructure is now being targeted at the same speed as traditional web vulnerabilities.
Bitwarden CLI Compromised in npm Supply Chain Attack — Rotate Credentials If You Installed v2026.4.0
Attackers hijacked Bitwarden's npm publishing pipeline for 90 minutes on April 22, injecting a self-spreading credential-theft worm into @bitwarden/cli@2026.4.0. The malicious version has been removed, but anyone who installed it must rotate all secrets immediately.
Page 39 of 61 · 549 articles