The 10 biggest things in tech, Thu, Jun 11

The ten biggest stories in tech today, scored 1–10 and ranked from 150+ engineering sources. Free, updated every morning.

  1. Anthropic Fable 5 launch controversy

    Anthropic launched Fable 5 (its first Mythos-class frontier model) to immediate backlash over aggressive guardrails rejecting innocuous prompts, undisclosed safety restrictions, and high token burn rates — forcing a public apology and policy reversal within days.

    #1 · ai · scored 7/10

  2. DiffusionGemma open-weight release

    Google DeepMind released DiffusionGemma-26B under Apache 2.0, achieving 857 tokens/second (4x faster than standard Gemma) and making the model available on NVIDIA Build and Hugging Face, marking a notable leap in open-weight on-device inference efficiency.

    #2 · ai · scored 7/10

  3. OpenAI models land on AWS Bedrock

    GPT-5.5, GPT-5.4, and Codex reached general availability on Amazon Bedrock — including AWS GovCloud — following a revised Azure exclusivity arrangement, with Codex shifting to pay-per-token billing.

    #3 · ai · scored 7/10

  4. German court limits AI search in EU

    A German court ruled Google AI Overview violates EU competition and consumer protection laws, setting a precedent that could restrict AI-powered search deployments across Europe.

    #4 · ai · scored 7/10

  5. npm v12 supply-chain security overhaul

    npm v12 (July 2026) makes install scripts opt-in and disables remote URL dependencies by default, with GitHub explicitly framing the breaking changes as a direct counter to supply-chain attacks abusing npm lifecycle hooks.

    #5 · security · scored 7/10

  6. ShinyHunters breach spree

    ShinyHunters claimed breaches of 100+ Oracle PeopleSoft deployments across enterprises and universities and separately exfiltrated 40 GB of student records from Nottingham University, establishing the group as a prolific enterprise threat actor this week.

    #6 · security · scored 7/10

  7. AI compresses vulnerability weaponization timeline

    Multiple reports — covering AI-assisted phishing, leaked worm code, Claude vulnerability patches, and a CISO analysis — converge on a single narrative: AI has eliminated the historical delay between vulnerability discovery and weaponization, forcing enterprises to rethink budget allocation toward breach simulation over traditional patch management.

    #7 · security · scored 7/10

  8. North Korean IT worker infiltration campaign

    CrowdStrike attributes roughly 50% of attacks on US, European, and Asian tech firms over the past 12 months to North Korean actors posing as remote IT workers and recruiters, marking the campaign as one of the most persistent supply-chain threats currently active.

    #8 · security · scored 7/10

  9. Windows zero-day publicly disclosed

    A security researcher publicly disclosed a new Windows zero-day, citing grievances over Microsoft's handling of previous vulnerability reports, raising concerns about researcher-vendor relations driving weaponizable disclosures.

    #9 · security · scored 7/10

  10. Coupang $400M privacy enforcement record

    South Korea fined Coupang over $400M for a data breach affecting 30M+ customers, setting a record regulatory penalty that signals aggressive enforcement posture for platform-scale privacy violations.

    #10 · security · scored 7/10