GEEK HAUS
← All dates

Daily reads: Jun 30, 2026

12 reads selected from 14 stories across 5 sources published in this period.

We take turns across sources, starting with their most recent stories. This is a reading selection, not an importance ranking or an archive of the email newsletter.

Dates follow Korea Standard Time. Weekly selections cover Monday through Sunday.

Our editorial policy and selection method
  1. 01 / The Verge

    T-Mobile is booting customers from its oldest plans

    T-Mobile has started notifying some customers that it will retire older legacy plans, including plans dating back to the 3G and 4G eras. Chief marketing officer Allan Samson confirmed the move, saying affected subscribers will be transitioned to modern plans built for the carrier’s 5G network.

    Why it matters

    The shift shows carriers are increasingly willing to phase out older, often cheaper plans as network technology and pricing strategies evolve. For customers, the change could mean better plan features but also raises concerns about higher costs and reduced flexibility for longtime subscribers.

  2. 02 / TechCrunch

    Chamath Palihapitiya raises $135M Series A for his AI coding startup, takes CEO role

    8090 Labs, an AI coding startup founded by Chamath Palihapitiya, raised a $135 million Series A led by Salesforce Ventures with backing from WndrCo, Craft Ventures, and other prominent investors. The company’s Software Factory product targets corporate engineering teams, promising production-quality software development with enterprise controls like audit trails.

    Why it matters

    The funding underscores investor demand for AI coding tools built for large companies rather than individual developers or prototype workflows. Palihapitiya taking the CEO role signals that enterprise AI agents are becoming a major battleground for venture-backed startups seeking to turn generative AI into core software infrastructure.

  3. 03 / VentureBeat

    DeepSeek open sources DSpark, a new framework to speed up LLM inference by up to 85%

    DeepSeek released DSpark, an open-source framework designed to accelerate LLM responses through speculative decoding without changing the model’s intended output. The company also published a technical paper, model checkpoints, and DeepSpec tooling under the MIT license, making the system broadly usable for research and commercial deployment.

    Why it matters

    DSpark targets one of AI’s biggest deployment bottlenecks: the cost and latency of serving large models at scale. Its release reinforces a broader trend toward open infrastructure that can narrow the performance gap between proprietary and open AI systems, especially for chatbots, coding tools, and agentic workflows.

  4. 04 / hccf.onmy.cloud

    Human-Centered Computing Foundation launches campaign for .self, a top-level domain for ethical self-hosted web services

    The Human-Centered Computing Foundation is seeking to secure .self as a new top-level domain focused on ethical, human-centered technology. The group says the domain would support an alternative web architecture that reduces dependence on data-extractive platforms.

    Why it matters

    The proposal reflects growing interest in digital sovereignty and self-hosted infrastructure as users and developers push back against centralized web platforms. If approved, .self could become a branding and discovery layer for privacy-focused services, though adoption would depend on tooling, governance, and community trust.

  5. 05 / quesma.com

    Qwen 3.6 27B impresses as a practical local AI model for coding, writing, and everyday development tasks

    Piotr Migdał says Qwen 3.6 27B is the first local model he has found compelling as a general-purpose assistant, despite running hot and slower than its MoE sibling. In tests, it handled constrained writing, poetry, and a one-prompt hexagonal Minesweeper app with better instruction-following than Qwen 3.6 35B A3B.

    Why it matters

    The article reflects a broader shift toward local AI becoming useful for real developer workflows rather than just demos. If models around this size can reliably follow instructions and generate working projects on consumer hardware, they could reduce reliance on cloud LLMs for privacy-sensitive or cost-conscious development.

  6. 06 / The Verge

    After a great start, DC’s new cinematic universe is already slowing down

    After James Gunn’s Superman raised expectations for Warner Bros. Discovery’s revamped DC universe, the studio’s plan to quickly expand with Supergirl already looked risky. The article says Supergirl has landed poorly with critics, casting doubt on whether the new DCU can maintain its early excitement.

    Why it matters

    The setback underscores how fragile franchise momentum can be, even after a well-received reboot. Warner Bros. Discovery needs each early DCU entry to build trust, and a weak follow-up could make audiences more skeptical of its broader interconnected strategy.

  7. 07 / TechCrunch

    Gemini’s personalized AI image generation is now free for US users

    Google is expanding Gemini’s personalized AI image generation to eligible free users in the US. The feature can create images informed by a user’s interests and data from connected Google apps.

    Why it matters

    The rollout shows Google pushing personalized AI features beyond paid tiers to broaden adoption. Tying image generation to account data could make outputs more relevant, while also raising the stakes for privacy controls and user trust.

  8. 08 / VentureBeat

    The attack that hijacked Claude Code came through Sentry. Datadog, PagerDuty, and Jira have the same exposure.

    Tenet Security says a crafted Sentry error event can inject attacker instructions into diagnostic data that Claude Code, Cursor, and Codex may treat as trusted input. In controlled tests across 100-plus targets, the agentjacking technique succeeded 85% of the time, while common defenses such as EDR, WAF, IAM, and firewalls did not flag it.

    Why it matters

    The finding highlights a new security gap where trusted observability and workflow tools become attack paths once connected to agents that can run commands. As AI coding agents gain deeper access to developer environments, organizations will need controls that limit what agents can execute from external data, not just protect credentials or perimeters.

  9. 09 / The Verge

    Leaked iPhone 18 Pro photos reportedly wound up on the dark web

    Leaked images and parts lists allegedly tied to the iPhone 18 Pro appeared on the dark web after a breach at Apple supplier Tata Electronics, Reuters reports. The images reportedly show a drop test unit with a three-camera layout and Apple logo, while researchers say ransomware group World Leaks posted more than 200,000 related files.

    Why it matters

    The incident highlights how Apple’s expanding supply chain can create new exposure points for unreleased hardware details. As manufacturing shifts and diversifies, supplier security is becoming a more important part of protecting product secrecy and brand control.

  10. 10 / TechCrunch

    Watch out, Amazon: The Kobo eReader now has a Goodreads rival

    Kobo users can now automatically sync their reading progress with StoryGraph, helping them track books, stats, and reading challenges more easily. The integration gives Kobo owners a built-in option beyond Amazon-owned Goodreads.

    Why it matters

    The move reflects growing demand for independent reading platforms as users seek alternatives to Amazon’s book ecosystem. For Kobo, tighter StoryGraph support could make its eReaders more appealing to readers who value richer tracking without relying on Goodreads.

  11. 11 / The Verge

    Tidal won’t pay royalties on AI-generated music but isn’t banning it outright

    Tidal says it will begin labeling tracks it identifies as 100% AI-generated on July 15, rather than banning them from the service. Starting immediately, those tracks will no longer earn royalties, as the company says payments should go to music directly created, written, and performed by people.

    Why it matters

    The policy reflects a growing industry shift toward AI disclosure instead of blanket removal, as platforms try to balance experimentation with artist protections. Tidal’s royalty cutoff could pressure other streaming services to define how they detect and monetize synthetic music.

  12. 12 / TechCrunch

    Waymo and Uber quietly part ways in Phoenix

    Waymo and Uber have quietly ended their autonomous ride-hailing partnership in Phoenix. Uber says it is preparing to launch a separate autonomous vehicle partnership in the city, but has not identified the new partner.

    Why it matters

    The split suggests Uber is continuing to diversify its robotaxi strategy rather than relying on a single provider. Phoenix remains a key testing ground for robotaxi commercialization, and the unnamed partner could signal shifting alliances in the autonomous mobility market.