AI starts quoting itself, Google retries Glass, and Bill Gates taps the brakes
Hi from Hiswai!
Last week in tech was a reminder that momentum does not always mean maturity. ChatGPT was caught leaning on other AI generated sources, raising uncomfortable questions about where truth comes from when machines learn from machines. Google dusted off smart glasses with a quieter, more wearable pitch, while Bill Gates stepped in to cool an overheated investment narrative. At the same time, infrastructure concerns moved closer to home as data centers met local power grids, and messaging apps quietly became new distribution channels. Together, these stories show technology stretching into new territory while still grappling with old constraints. Below are the updates that defined that tension.
ChatGPT starts citing other AIs and the loop gets messy
A new report found that ChatGPT's latest model has been pulling information from Grokipedia, an encyclopedia generated entirely by another AI system. In cases involving obscure topics, the chatbot referenced content that has no direct human authorship or editorial oversight, creating a recursive loop where AI systems reinforce each other's outputs. This raises fresh concerns about information quality, especially as hallucination issues remain unresolved and bad actors increasingly flood the web with synthetic content designed to influence models.
Why it matters: When AI starts learning from AI, errors can compound quickly and quietly. Trust in AI depends on grounding, not repetition. If models begin citing synthetic sources as authority, misinformation risks scaling faster than traditional fact checking can keep up, turning uncertainty into digital folklore that looks credible simply because it is repeated.
Google gives smart glasses another try and bets on better timing
Google is returning to smart glasses in 2026 with two new AI powered designs, reopening a chapter it first tried to write more than a decade ago. After earlier versions struggled with privacy backlash and social awkwardness, the new approach focuses on cleaner design, audio first experiences, and tighter AI integration. One version includes a display while another skips screens entirely, reflecting lessons learned from competitors who prioritized wearability over novelty.
Why it matters: Wearables only work when people want to wear them. If Google can align AI utility with social acceptance, smart glasses could finally move from curiosity to everyday tech. If not, it will be another reminder that timing and design matter as much as capability.
Bill Gates cools the AI investment frenzy
Bill Gates is urging investors to temper expectations as money continues to flood into AI. While calling the technology deeply profound, he warned that the market will be hypercompetitive and that many companies carrying sky high valuations will not survive. Gates pointed to massive infrastructure spending by hyperscalers and the growing gap between real revenue and speculative pricing, especially among unprofitable startups commanding extraordinary private valuations.
Why it matters: His message is not anti AI, but a reminder that transformative technology does not guarantee universal winners. AI may be reshaping the economy, but markets still obey gravity. As capital concentrates and competition intensifies, valuation discipline will matter more than narrative, separating durable businesses from hype driven bets.
OpenAI tries to keep its data centers from spiking local power bills
OpenAI says it will fund new energy infrastructure to prevent its massive Stargate data center buildout from raising electricity costs for nearby communities. The company plans to create custom agreements for each site, working with regulators and residents to add generation and storage capacity rather than leaning on existing grids. The pledge comes as concern grows that large scale AI facilities could dramatically increase power demand over the next decade.
Why it matters: AI infrastructure is becoming a public issue, not just a corporate one. If companies want to scale without backlash, they will need to prove that growth does not come at the expense of local costs, trust, or energy stability.
Google turns shared links into instant previews inside Messages
Google has added a new Messages bot that generates link previews directly inside chat threads when users share URLs. The fetcher works only when triggered by a user action, pulling page content to create visual previews rather than crawling the web on its own schedule. It joins a growing set of Google systems designed to respond on demand, expanding how content shows up across chat and productivity tools. For site owners, the traffic is identifiable through a dedicated user agent and published IP ranges.
Why it matters: Messaging apps are becoming distribution channels, not just communication tools. As links turn into previews at the moment of sharing, visibility and context shift from search pages to private conversations, changing where attention is captured and how content travels.
Enjoyed this newsletter?
Forward it to friends and invite them to join our growing Tech community by subscribing below!
Explore our dashboards to find more stories like these.
You can also create an account on Hiswai and start receiving updates on any topic of your interest in real-time.
That wraps up this week's tech brief. What stands out is how growth is starting to meet resistance from reality, whether that is physics, economics, or credibility itself. As platforms scale, the pressure shifts from moving fast to getting things right, from novelty to trust, and from ambition to accountability. Those fault lines are where the next big moves will happen. We will be back next week with a clear read on what continues to hold and what starts to crack.
Cheers,
