Eyes on the Chaos
Thursday, May 7, 2026

Archived edition

Thursday, May 7, 2026

9 stories curated from 16 sources

In today's issue

DesignEthicsProduct
  1. 01
    Mira Murati tells the court that she couldn't trust Sam Altman's words

    OpenAI's former CTO testified that CEO Sam Altman lied about safety standards.

  2. 02
    Using AI for Just 10 Minutes Might Make You Lazy and Dumb, Study Shows

    Brief AI assistant use may impair people's thinking and problem-solving abilities.

  3. 03
    Chrome's AI features may be hogging 4GB of your computer storage

    Chrome automatically downloads large AI model files without clear user consent.

  4. 04
    Google's AI search summaries will now quote Reddit

    Google AI search now surfaces perspectives from Reddit and social forums.

  5. 05
    You are not immune to mode collapse

    Digital experiences risk becoming homogenized and losing authenticity through AI.

  6. 06
    DeepSeek could hit $45B valuation from its first investment round

    Chinese AI lab achieves massive valuation after training models cost-effectively.

  7. 07
    Tinder owner Match Group is slowing hiring to pay for its increased use of AI tools

    Match Group cutting hiring plans because AI tools are expensive to operate.

  8. 08
    SpaceX may spend up to $119B on 'Terafab' chip factory in Texas

    SpaceX planning massive semiconductor facility for vertically integrated chip manufacturing.

  9. 09
    An Interview with Joanna Stern About Living With AI

    Tech journalist discusses practical AI adoption and launching independent media company.

AI Research & News

Mira Murati tells the court that she couldn't trust Sam Altman's words

The Verge

EthicsProduct

OpenAI's former CTO testified that CEO Sam Altman lied about safety standards.

  • Key allegation: Murati testified under oath that Altman falsely claimed OpenAI's legal department cleared a new AI model to bypass the deployment safety board.
  • Trial context: This testimony is part of the ongoing Musk v. Altman lawsuit over OpenAI's mission and governance.
  • Safety implications: The testimony suggests internal conflicts over AI safety processes at one of the industry's leading companies.
  • Leadership credibility: A former top executive publicly stating the CEO lied raises serious questions about OpenAI's internal culture and decision-making.

For product

Consider how safety review processes are documented and communicated in your own AI product development — verbal assurances from leadership may not be sufficient for high-stakes decisions.

Using AI for Just 10 Minutes Might Make You Lazy and Dumb, Study Shows

Wired

DesignEthicsProduct

Brief AI assistant use may impair people's thinking and problem-solving abilities.

  • Research findings: New study shows that just 10 minutes of AI assistant usage can negatively impact users' cognitive abilities and problem-solving skills.
  • Dependency risk: The research suggests people quickly develop reliance on AI tools, potentially atrophying their own thinking processes.
  • Speed of impact: The cognitive effects appear remarkably quickly, raising questions about long-term consequences of AI integration in workflows.
  • Design challenge: Results highlight the need to balance AI assistance with maintaining users' cognitive engagement and skill development.

For design

When designing AI-assisted workflows, build in friction points that require human reasoning rather than defaulting to fully automated solutions.

Chrome's AI features may be hogging 4GB of your computer storage

The Verge

DesignProduct

Chrome automatically downloads large AI model files without clear user consent.

  • Storage impact: Chrome is automatically installing a 4GB weights.bin file for Gemini Nano AI model when certain features are enabled.
  • User awareness: Many users are discovering unexpected storage drops without realizing Chrome's AI features were the cause.
  • Consent question: The automatic download raises questions about user consent for resource-intensive AI feature installations.
  • Precedent concern: As more browsers integrate AI, this could become a widespread issue affecting device performance and user experience.

For product

Establish clear consent flows and resource disclosure for any AI features that require significant local storage or processing power.

Product & UX

Google's AI search summaries will now quote Reddit

The Verge

DesignProduct

Google AI search now surfaces perspectives from Reddit and social forums.

  • Source expansion: Google's AI search features now include 'a preview of perspectives' from Reddit, social media, and web forums alongside traditional results.
  • Human-first approach: The update aims to connect search queries with real human conversations and firsthand experiences on similar topics.
  • Trust signals: Google is positioning this as helping users find information from sources they 'know and trust' rather than generic web content.
  • Quality risk: While this could improve relevance for niche queries, it also introduces potential for surfacing low-quality or misleading forum discussions.

For design

Users increasingly expect AI-generated content to include human perspectives and social proof — consider how to surface authentic user voices in your own AI-powered features.

You are not immune to mode collapse

Sidebar.io

DesignProduct

Digital experiences risk becoming homogenized and losing authenticity through AI.

  • Mode collapse risk: Digital products face the danger of becoming increasingly similar and losing their unique characteristics as AI tools standardize design patterns.
  • Authenticity challenge: The widespread use of AI-generated content and design elements threatens to make digital experiences feel generic and artificially uniform.
  • Differentiation imperative: Companies need to actively work to maintain distinctive brand voices and user experiences in an AI-homogenized landscape.
  • Human touch: Preserving genuine human input and decision-making becomes crucial for keeping digital products feeling authentic and useful.

For design

Establish design principles that explicitly preserve your product's unique voice and avoid defaulting to AI-suggested patterns that may be overly generic.

Business & Strategy

DeepSeek could hit $45B valuation from its first investment round

TechCrunch

Product

Chinese AI lab achieves massive valuation after training models cost-effectively.

  • Valuation milestone: DeepSeek is reportedly seeking a $45 billion valuation in its first major funding round.
  • Cost advantage: The company gained attention for training large language models using a fraction of the compute power and cost of U.S. competitors like OpenAI.
  • Competitive threat: Their efficient training methods challenge the assumption that only massive capital investment can produce competitive AI models.
  • Geopolitical impact: Success demonstrates China's ability to compete in AI despite U.S. chip export restrictions.

For product

The cost-efficiency breakthrough suggests you may not need the most expensive AI models for your use cases — evaluate whether cheaper, more efficient alternatives can meet your product requirements.

Tinder owner Match Group is slowing hiring to pay for its increased use of AI tools

TechCrunch

Product

Match Group cutting hiring plans because AI tools are expensive to operate.

  • Resource tradeoff: Match Group is slowing hiring for the rest of the year specifically because AI tools 'cost a lot of money.'
  • Budget reality: The decision reveals that AI implementation costs are significant enough to impact fundamental business decisions like staffing.
  • Industry signal: This may be the first major public admission that AI operational costs are forcing companies to make difficult budget choices.
  • Strategic question: The move raises questions about whether AI productivity gains are offsetting their operational costs in the near term.

For product

Build detailed cost models for AI feature rollouts that account for ongoing operational expenses, not just development costs — they may be higher than expected and require budget reallocation.

SpaceX may spend up to $119B on 'Terafab' chip factory in Texas

TechCrunch

SpaceX planning massive semiconductor facility for vertically integrated chip manufacturing.

  • Massive investment: SpaceX is proposing to spend up to $119 billion on a 'multi-phase, next-generation, vertically integrated semiconductor manufacturing and advanced computing fabrication facility' in Texas.
  • Strategic shift: The move represents SpaceX expanding beyond aerospace into critical chip manufacturing infrastructure.
  • Supply chain control: Vertical integration would give SpaceX direct control over semiconductor supply chains for its space and AI ambitions.
  • Scale comparison: The proposed investment dwarfs most semiconductor manufacturing projects and signals massive long-term chip demand expectations.

For product

The scale of investment suggests compute-intensive AI applications will require fundamentally different chip architectures — consider how this might affect your long-term technical roadmap.

An Interview with Joanna Stern About Living With AI

Stratechery

Tech journalist discusses practical AI adoption and launching independent media company.

  • Practical AI use: Stern discusses real-world experiences with AI integration in daily life and work, providing grounded perspective beyond tech hype.
  • Media independence: The interview covers her decision to start her own media company, reflecting broader trends in tech journalism and content creation.
  • AI adoption reality: Her insights offer a pragmatic view of how AI tools actually perform in professional workflows versus marketing promises.
  • Industry credibility: As an established tech journalist, her perspective carries weight for understanding AI's practical business impact.

For product

Seek out grounded, real-world usage data rather than relying on promotional case studies when evaluating AI tools for your team workflows.