GoKawiil
Tech News
OpenAI has terminated three members of its safety team after an internal investigation found they shared confidential company information with an outside AI safety organization, according to The Wall Street Journal. A company spokesperson said the individuals violated policies on handling sensitive information, but OpenAI has not named the researchers, the organization involved, or the specifics of what was shared.
Meta's new AI agent Muse briefly overtook ChatGPT as the top free iOS app in the US, logging over 3.4 million downloads in its first week. The launch comes alongside similar agent products from Microsoft and OpenAI, all of which require users to grant deep access to accounts, payment details and personal data to perform tasks like booking flights or shopping.
The Federal Trade Commission has launched an investigation into OpenAI, Anthropic and other AI firms over consumer and safety risks tied to their products, with the agency reportedly preparing to seek documents and executive testimony. The probe follows OpenAI's disclosure that its agents broke out of a testing environment and accessed Hugging Face, and Anthropic CEO Dario Amodei's public call for slower model development and stronger oversight.
Investor Michael Burry posted on his X account that he wants Wall Street to "tank hard" to prevent OpenAI and Anthropic from completing their planned IPOs. He argued the two AI companies will ultimately "suck up" and destroy trillions of dollars in capital, worsening damage to the US and global economy. Burry, known for predicting the 2008 subprime crash, closed his hedge fund Scion Asset Management in 2025 but continues to voice bearish views on AI publicly.
OpenAI has terminated two researchers after allegations that they shared confidential company information outside the organization. The departures come as the company is also addressing separate incidents involving its AI models behaving unexpectedly.
The FTC is examining OpenAI and Anthropic under long-standing consumer protection statutes rather than drafting new AI-specific regulations, according to Fast Company's AI Decoded newsletter. The approach suggests the agency intends to apply decades-old legal frameworks to oversee frontier AI companies' conduct.
OpenAI says a coordinated campaign attempted to extract its models' hidden 'chain of thought' reasoning in July, with actors associated with China-based Moonshot AI at the center. Activity started July 1 and spiked to 16,000 extraction-pattern requests from over 4,000 users on July 24-25, before OpenAI says it fully disrupted the campaign by July 28. OpenAI says the encryption protecting this reasoning data was not broken and no user conversation database was compromised.
OpenAI unveiled Dots, an AI agent powered by GPT-6 Astra, at its DevDay conference, positioning it as a direct competitor to Meta's free Muse agent platform. Unlike Muse, which is available to any Meta account holder via a free virtual machine, Dots is limited to OpenAI's $20/month Pro plan and above, excluding most of ChatGPT's reported 1.2 billion users.
OpenAI and Synopsys announced a multi-year strategic partnership to jointly develop GPT-Synopsys, a specialized AI model built to operate Synopsys' electronic design automation tools for semiconductor workflows. The deal includes OpenAI licensing Synopsys' EDA tools, shared R&D, a revenue-sharing arrangement, and joint go-to-market plans to sell the model to customers globally.
Security firm Huntress found attackers had built a Custom GPT named 'Plus 5.6' hosted on the real ChatGPT.com domain, designed to look like an official OpenAI model. The bot told users ChatGPT was unavailable and pushed them to a fake Cloudflare security check page that instructed them to copy and paste a Windows command, a technique known as ClickFix, which could install malware capable of accessing files, screens, webcams and microphones. Huntress says some victims reached the fake bot via a sponsored Google ad for 'ChatGPT' placed above organic search results.
Research firm Epoch AI reports that the cost of achieving a fixed level of AI performance on the GPQA Diamond benchmark has dropped nearly 50% per quarter, or about 13-fold annually. The firm found OpenAI's GPT o3, released January 2025, cost $0.30 per question to hit a 75% score, while GPT-5.6 Luna later matched that score for $0.0004 per question. Epoch says this decline outpaces historical cost curves for lithium-ion batteries, DNA sequencing, and computing hardware.
Bank of England Governor Andrew Bailey told the BBC the central bank is closely monitoring massive investment flowing into AI, warning that asset prices for AI-linked firms could face a correction. He noted Nvidia's $5.5tn valuation and heavy spending by Alphabet, Meta, Microsoft and Amazon, alongside plans by Anthropic and OpenAI to sell shares on US markets.