“GPT-4 Kinda Sucks”

Sam Altman on GPT-4, “It Kinda Sucks”​

Sam Altman, CEO of OpenAI, shockingly admits GPT-4 “kinda sucks” despite its impressive capabilities, already looking ahead to the next potential leap with GPT-5.

• The CEO of OpenAI, Sam Altman, surprisingly stated that GPT-4 “kinda sucks” in an interview with renowned American computer scientist Lex Fridman.
• Despite the impressive capabilities of GPT-4, Altman is more excited about the potential of GPT-5 or whatever their next model will be called (he doesn’t know what to name it yet).
• He believes the current tools, including GPT-4, will seem inadequate in the future, and it will be OpenAI’s responsibility to push the boundaries further. Altman acknowledged that GPT-4 has a “glimmer of something amazing” but is often best used as a brainstorming partner rather than for complex, multi-step problems.
• While GPT-4 has propelled OpenAI to 100 million weekly users and an $80 billion valuation, Altman is already looking ahead to the next iteration, which he hopes will be a significant leap forward, similar to the jump from GPT-3 to GPT-4.

Google Gemini + iPhone​

In a potential industry-shaking deal, Apple is in talks with Google to integrate its powerful Gemini AI models into future iPhone software.

• Apple is in active negotiations with Google to license and integrate Google’s Gemini generative AI models into upcoming iPhone software features. This potential deal would allow Apple to power new AI capabilities on the iPhone using Google’s advanced language models.
• The talks between the two tech giants are setting the stage for a major agreement that could shake up the AI industry. Apple has recently engaged in talks with OpenAI regarding the possibility of utilizing its AI model alongside or instead of Google’s Gemini.
• If the deal with Google goes through, it would build upon the existing search partnership between the two companies. Integrating Gemini into iOS would bring powerful generative AI tools to hundreds of millions of iPhone users worldwide.
• While the negotiations are still ongoing, the potential Apple-Google AI deal highlights the demand for cutting-edge generative AI technology across consumer products.

Nvidia’s Absolute Monster GPU​

Nvidia CEO Jensen Huang revealed the company’s powerful new Blackwell GPU and specialized hardware that would accelerate the computing power of AI models.

• Nvidia CEO Jensen Huang delivered a highly technical keynote at the company’s GTC (GPU Technology Conference). Unlike typical consumer tech keynotes, Huang “nerded out” for two hours on the deep technical details and capabilities of Nvidia’s new products and platforms.
• The keynote highlighted Nvidia’s return to its roots at the intersection of computer graphics, physics, and AI. Huang announced the new Blackwell platform, described as an “absolute monster” graphics processing unit (GPU) that will enable higher resolutions, better quality, and higher frame rates for gaming and other graphics-intensive applications.
• Rather than a slick consumer-focused presentation, the keynote was an unapologetically developer-centric event filled with technical jargon, acronyms, and in-depth discussions of algorithms, architectures, and mathematics behind Nvidia’s latest innovations.
• Huang emphasized the need for bigger and more powerful GPUs to push the boundaries of accelerated computing across graphics, AI, and scientific domains. The keynote showcased Nvidia’s core mission to specialize hardware for performance beyond general-purpose computing.