OpenAI & Google’s AI Search Wars

Google’s recent “Learn About” AI tool, currently available within the U.S., seeks to streamline knowledge extraction through interactive features like timelines, quizzes, and “Go Deeper” or “Simplify” options. Users can attach images to queries, making it a dynamic alternative to straightforward search.

Unlike Google’s traditional AI-powered overviews, which might overwhelm with varied formats, Learn About is designed for focused, conversational learning. Although still experimental, it’s seen as a possible rival to Perplexity, which positions itself as a go-to tool for deep knowledge discovery.

Positioned as a possible competitor to Perplexity, which recently secured major funding, Google’s tool stays experimental and limited in functionality—users cannot resume conversations once they close the tab. Read more.

Google’s Gemini API is riding a wave of growth, with usage skyrocketing fourteen-fold in only six months, in accordance with CEO Sundar Pichai. This API is now embedded in key services like Maps and Search, potentially reaching nearly two billion users—over a billion of whom engage via Search. Meanwhile, Google Cloud is having fun with a 35% revenue boost, hitting $11.4 billion, due to the AI surge.

Pichai also teased upcoming AI initiatives, including Project Astra, a multimodal assistant able to processing text, video, and audio in real time. Expected releases include Gemini 2.0 and an online navigation agent named “Jarvis” in December 2024. AI now accounts for over 25% of latest code generated at Google, reflecting significant operational efficiencies much like those reported by Amazon. Read more.

The race for AI-powered search dominance heats up as Google, OpenAI, and Perplexity each aim to reshape how we discover information.

Just hours before OpenAI’s ChatGPT Search went live, Google enabled its Gemini AI platform to access real-time Google Search data, debuting its “Grounding with Google Search” feature. This integration allows developers to embed up-to-the-minute, citation-backed information into applications, leveraging Google’s vast search infrastructure to prioritize high-value queries, priced at $35 per 1,000 queries. The timing underscores Google’s intent to leverage its search pedigree in AI applications, catering especially to developers constructing search-centric tools.

OpenAI’s ChatGPT Search, meanwhile, has rolled out with a give attention to consumer accessibility, delivering conversational, source-attributed responses across topics like news and stocks, with plans to achieve free users soon. Built on GPT-4o, it builds on OpenAI’s earlier SearchGPT trial and taps into partnerships with news organizations just like the Associated Press and Vox Media, providing verified, premium content.

Perplexity also competes on this space by offering model flexibility—allowing users to make a choice from Claude, ChatGPT, and Sonar—often providing more structured, in-depth answers.

As these platforms redefine search, each faces the challenge of balancing user engagement, cost, and accuracy, with implications for licensing costs and the sustainability of AI-enhanced search in a crowded field. Read more.

In a recent Reddit AMA, OpenAI’s Sam Altman and team clarified that GPT-5 won’t debut this 12 months; as a substitute, they’ll refine the o1 series, which can eventually merge with GPT. They’re enhancing context windows for ChatGPT Plus and GPT-4o to higher handle complex tasks. A brand new text-to-image model is on the best way, together with image input support within the o1 series.

On the AGI front, they’re training with specialized medical literature while ensuring HIPAA compliance. The discussion highlighted the importance of open-source models, user control over NSFW content, and ongoing efforts to scale back hallucinations in AI outputs, with multilingual capabilities receiving a lift in the newest models. Read more.