If you’re better at critiquing images than creating them, ChatGPT’s improved image generation is worth a try. Recent enhancements let you direct iterative changes, fix specific areas, and finally get text that’s spelled correctly.
OpenAI has traced ChatGPT’s bizarre goblin fixation to training gone awry, but the creatures keep escaping—as Adam Engst discovered when goblins popped up in a conversation about a conference presentation.
Apple and Google have announced a multi-year partnership to base the next generation of Apple’s foundation models on Google’s Gemini technology. This collaboration will drive future Apple Intelligence features, including an improved Siri expected later this year.
What started as a simple need—digitizing training pace tables for a workout app—became a test of OCR tools. ChatGPT Atlas won where others failed, autonomously processing five photos into perfect CSV data despite page curl and dense columns.
Anthropic is prompting Claude users to opt into having their conversations used for training future models. While the immediate privacy risks seem low, the long-term implications of AI training data remain unclear, suggesting a cautious approach.
The AI industry is standardizing around $20 monthly subscriptions for premium features. However, with multiple services requiring separate subscriptions and third-party apps dependent on model providers, will this approach ultimately prove practical for both consumers and developers?
With its new study mode, ChatGPT becomes less of an answerbot and more of an interactive tutor, helping users understand concepts rather than just giving them answers.
Our Do You Use It? poll reveals minimal engagement with most Apple Intelligence features, suggesting these AI-powered capabilities aren’t yet meeting user needs.
A detailed analysis demonstrates that individual ChatGPT use has minimal environmental impact compared to other everyday activities.
Looking for better search results? Adam Engst explores how AI-powered answer engines are transforming online research by providing direct answers rather than just links.
On the Chit Chat Across the Pond podcast, Adam Engst discusses the accuracy of AI-powered transcription and suggests that AI-powered search engines represent a fundamental shift in how we access and interact with information.
Paul Kafasis relates the hilariously incorrect results of asking Siri to identify Super Bowl winners, which leads John Gruber to compare Siri with other answer engines, a comparison in which Siri comes out smelling like old socks.
Adam Engst explores how ChatGPT can transform complex spreadsheet data into valuable visualizations, sharing insights from analyzing his running club’s registration patterns and promotional channel analysis.
In light of Apple Intelligence’s notification missteps, Adam Engst examines the value of summarization, overcoming an initial skepticism to argue that summaries are intrinsic to human expression but should be used judiciously.
Despite Apple's efforts to improve Siri with ChatGPT integration, Adam Engst finds that the new feature presents more challenges than solutions. Functional hiccups and broader concerns cast doubt on its effectiveness.