- Pivot 5
- Posts
- OpenAI’s SearchGPT prototype
OpenAI’s SearchGPT prototype
1. OpenAI’s SearchGPT prototype
We’re testing SearchGPT, a temporary prototype of new AI search features that give you fast and timely answers with clear and relevant sources.
We’re launching with a small group of users for feedback and plan to integrate the experience into ChatGPT. openai.com/index/searchgp…
— OpenAI (@OpenAI)
6:09 PM • Jul 25, 2024
OpenAI is testing SearchGPT, a prototype of new search features that combines AI models with web information to provide fast, timely answers. The prototype is launching to a small group of users and publishers for feedback.
The goal is to enhance conversational capabilities and provide clear links to relevant sources. OpenAI is also partnering with publishers and creators to help users discover publisher sites and experiences. The prototype is separate from training OpenAI's generative AI foundation models.
Read the full story here
2. Websites accuse AI startup Anthropic of bypassing their anti-scraping rules and protocol
Freelancer and iFixit CEO Kyle Wiens have accused AI startup Anthropic of ignoring its "do not crawl" robots.txt protocol to scrape their websites' data. Anthropic's ClaudeBot has been accused of 3.5 million visits from the company's crawler within four hours, and iFixit's servers were hit a million times in 24 hours.
AI firms like OpenAI have been the target of multiple lawsuits for copyright infringement, with publishers accusing them of copyright infringement. Wiens is open to discussing licensing content for commercial use with Anthropic.
Read the full story here
3. Elon Musk’s X under pressure from regulators over data harvesting for Grok AI
Elon Musk's X platform is under pressure from UK and Irish data regulators after users are consenting to their posts being used to build artificial intelligence systems via a default setting. The UK GDPR and EU data regulation prohibit companies from using "pre-ticked boxes" or "any other method of default consent".
The Data Protection Commission in Ireland has been speaking to X about data collection and AI models, and is "surprised" to learn of the default setting.
Read the full story here
4. Non-Google search engines blocked from showing recent Reddit results
Reddit's Content Policy has been updated to prevent crawling its site without explicit consent, resulting in minimal or no Reddit results from the past week.
This change has tightened Google's stranglehold on the search industry, as Google is the only search engine that can show recent Reddit results, despite concerns about SEO and AI spam farms, ads, and e-commerce links.
Read the full story here
5. Establishing Standards for Embodied AI
Embodied Artificial Intelligence involves embedding AI into robots, enabling them to perceive, learn, and interact with their environment. A standards framework is needed to ensure safety, reliability, performance, and usability.
Safety involves preventing accidents, reliability involves maintaining consistent performance, performance standards address task accuracy, and usability standards ensure user-friendliness.
Read the full story here
Advertise with Pivot 5 to reach influential minds & elevate your brand
Get your brand in front of 80,000+ businesses and professionals who rely on Pivot 5 for daily AI updates. Book future ad spots here.