Hey!

Got something exciting to share. Let’s dive in.⚡

🔥 It’s Going Viral:

Sundar Pichai Says No One Is Safe If the AI Bubble Pops

Google CEO Sundar Pichai has warned that if the AI bubble bursts, no company — not even Google — will be fully protected. He said today’s AI investments include “elements of irrationality,” drawing parallels to the dot-com era. Pichai also reminded users not to trust AI chatbots blindly, as the technology can still make mistakes.

Key Points:

• Pichai says the AI market could face a crash similar to the 2000 dot-com bubble.

• Google isn’t immune, though its full-stack approach may help it survive the impact.

• He urges users to verify critical information and not rely solely on AI responses.

Perplexity Set to Launch Free AI Shopping Tool

Perplexity is reportedly preparing to launch a free agentic shopping tool next week, making in-app purchases available to all users. This move follows similar shopping features from OpenAI and Google, and expands on Perplexity’s existing “Buy with Pro” option. The company also announced a partnership with PayPal, allowing users to search, shop, and pay directly through Perplexity.

Key Points:

• New tool will let users make purchases straight from answers, similar to Buy with Pro.

• Launch expected next week in the US, alongside PayPal integration for payments.

• Follows recent launches from OpenAI and Google in the AI shopping space.

Perplexity, Anthropic May Have Leaked Secrets on GitHub

A new report by cloud security firm Wiz claims that major AI companies, including Perplexity and Anthropic, may have accidentally exposed sensitive data on GitHub. The report found that about 65% of leading AI firms had exposure risks involving proprietary model details, API keys, and other credentials. Researchers warned that such leaks could reveal valuable information about the companies’ AI models and training processes, urging firms to adopt stronger monitoring tools.

Key Points:

• Wiz analysed companies from Forbes’ AI 50 list, including Anthropic, Mistral, Cohere, Midjourney, Perplexity, and others.

• Exposed data included API keys, model weights, dataset info, and tokens.

• Leaks occurred when developers’ GitHub repositories contained sensitive code or files.

🚀 New in Geekflare Connect

Now you can set default LLM models for your entire teams in Geekflare Connect.

🚀 Cut through the noise!

We’ve launched a WhatsApp channel to deliver clean, curated updates on AI, SaaS, and tech—no distractions, just the news you need.

📈 Trending in Tech:

Cheers,

Keval, Editor