Benchmark bet on Cerebras despite some early doubts
LinkedIn (https://www.linkedin.com/news/story/benchmark-bet-on-cerebras-despite-some-early-doubts-8838874/) 17 May 2026
Takeaways
- Benchmark recognized the potential of Cerebras’ innovative AI chips.
- Cerebras has since attracted major clients like OpenAI and AWS.
- Benchmark’s investment, initially around $270 million, is now worth billions.
Hindsight / Insight / Foresight
- Benchmark, an investor, holds a significant stake in Cerebras Systems.
Companies confront challenges of AI agent proliferation
LinkedIn (https://www.linkedin.com/news/story/companies-confront-challenges-of-ai-agent-proliferation-8853914/) 17 May 2026
Takeaways
- Several firms are trying to tackle the problems associated with “AI agent sprawl”—the rapid proliferation of AI-powered assistants—all without deterring companies from using AI technology.
- Risks of having too many AI agents include elevated computing bills and cybersecurity problems.
- Platforms like Anthropic’s Claude Cowork and OpenClaw make it easier for employees to create independent bots.
Hindsight / Insight / Foresight
- I’ve been warned against using OpenClaw due to its cybersecurity issues and risks.
Two-thirds of doctors are using AI for decision making
LinkedIn (https://www.linkedin.com/news/story/two-thirds-of-doctors-are-using-ai-for-decision-making-8803602/) 17 May 2026
Takeaways
- OpenEvidence, an AI-powered tool for clinical decision-making as well as quick, tailored medical insights, keeps gaining traction among doctors.
- Almost two-thirds of US doctors are active users along with another 1.2 million internationally.
Hindsight / Insight / Foresight
- Some physicians have raised concerns about OpenEvidence’s accuracy, potential impact on analytical thinking skills, and patient outcomes.
New AI-powered Smart Disputes will be enabled in June 2026
Stripe, email newsletter, 15 May 2026
Takeaways
- Smart Disputes will help users save time managing eligible disputes on card transactions and recover revenue—no action needed from users.
- It automatically compiles evidence and helps them add evidence to improve their chance of winning.
- Evidence is compiled using AI trained on real dispute data from the more than $1 trillion in payments Stripe processes annually.
Hindsight / Insight / Foresight
- If you win a dispute using Smart Disputes, you’ll pay 30% of the recovered disputed amount. If you don’t win, no fees for Smart Disputes apply.
- If you don’t want to use Smart Disputes, you can turn off auto-submission. If you opt out, you can still manually submit evidence generated by Smart Disputes.
The Shift From AI Assistant to AI Operator Has Started
Jonathan Mast, The AI Whisperer email newsletter, 15 May 2026
Takeaways
- AI is moving beyond chat into operational execution.
- Claude can now act directly inside existing business tools.
- The barrier to implementing AI operations is dropping fast.
- Early adopters will likely gain measurable efficiency advantages.
- Businesses already using Claude Pro or Team can test this immediately.
Hindsight / Insight / Foresight
- The businesses pulling ahead with AI won’t just be the ones asking better questions. They’ll be the ones building workflows where AI helps get work done.
- That shift is happening faster than most people realize.
New data shows exposure risks from AI vibe-coding apps
LinkedIn (https://www.linkedin.com/news/story/new-data-shows-exposure-risks-from-ai-vibe-coding-apps-7244420/) 11 May 2026
Takeaways
- New research highlights potential data vulnerabilities in applications built with vibe coding.
- A recent study of 380k publicly accessible vibe-coding assets found that 5k vibe-coded apps exposed corporate and personal information.
- Accessible sensitive data included financial records and patient conversations, raising regulatory concerns.
Hindsight / Insight / Foresight
- The findings highlight the weaknesses associated with shadow AI, as many organizations lack adequate access controls and strict governance policies.
Arizona students boo former Google CEO Eric Schmidt as he talks about AI during graduation speech
Business Insider (https://www.businessinsider.com/students-boo-eric-schmidt-google-ceo-ai-university-arizona-2026-5) 16 May 2026
Takeaways
- AI is transforming the global workforce, from how companies screen potential candidates to the types of skills companies are seeking. Despite young people increasingly adopting AI in their daily lives, surveys show they’re worried about what it means for their careers.
- The tech’s ability to automate many rote tasks has led some companies to cut back on entry-level hiring.
Hindsight / Insight / Foresight
- “The question is not whether AI will shape the world. It will. The question is whether you will have shaped artificial intelligence.” (Eric Schmidt)
- “AI is not likely to replace you. But someone using AI better than you might.” (Jensen Huang, Nvidia CEO)
The human labor that makes AI work
National Public Radio (NPR) (https://www.npr.org/2026/05/11/nx-s1-5816338/the-human-labor-that-makes-ai-work) 11 May 2026
Takeaways
- Even though it’s sold as an autonomous, inhuman tool, there’s a lot of human work that goes into it.
- While that work does happen in the US, many AI laborers also live overseas – and get paid less to do it. Everything that goes into making AI function – from the data it feeds on to the labor that trains it – smacks of colonialism, according to some scholars and researchers. This has led to the emergence of the term “data colonialism.”
Hindsight / Insight / Foresight
- “There are some nuances sometimes to the kind of work that can make [generative AI] less predatory. I would say the transparency would be one particular way [toward ethical generative AI]. Being open about the kind of work that is being done. Operating on consent, for example, is also another way of going about it. Creating on fair labor. Regulating the industry.” (Regine Sabato, freelance journalist based in the Philippines)