OWITH.ai - Only What's Important to Hear around AI and Tech

Good morning from OWITH.ai: the podcast that gives you only what's important to hear in AI and tech world.Significant advancements were made in AI this week, with Anthropic releasing the intelligent model Claude 3.7 Sonnet and OpenAI unveiling the GPT-4.5 language model. The University of Washington's Baker Lab also made strides in creating AI-designed proteins to neutralize snake venom toxins. Inception Labs introduced the Mercury language model, while the Gibberlink project showcased AI agents communicating through audio channels. These developments highlight the continuous progress in AI technology, impacting fields such as healthcare, language generation, and communication protocols.The limitations of large language models (LLMs) were discussed this week, focusing on challenges hindering them from reliable reasoning. New hybrid reasoning models like Claude 3.7 and Nvidia's Deepseek optimizations were introduced to address these limitations. The impact of AI on labor markets, economy, and advancements in reasoning models in 2025 were explored. The trade-off between response time and accuracy in reasoning-enhanced models was emphasized, along with insights on formal reasoning in AI.OpenAI faced significant challenges with the release of GPT 4.5, receiving harsh reviews questioning the company's promises. Despite struggles with costs, competitive advantage, and achieving AGI, OpenAI sees GPT-4.5 as a necessary step in scaling unsupervised learning. The focus on pre-training scaling laws over post-training reasoning capabilities signals a shift in AI development approach.While OpenAI launched GPT-4.5 as a strategic move towards AI advancements, challenges remain in staying ahead in the AI race. Criticism from Tulsi Gabbard towards UK government marks a shift in cybersecurity policy. Other updates include Meta launching a standalone AI app, Autodesk cutting its workforce, and Dell predicting a profit decline.President Trump's proposed trade tariffs impacted global stock markets significantly this week, leading to a decline in market indices. Analysts and financial institutions provided reactions and analysis on the market situation amidst other news stories involving various companies and CEOs.The impact of AI on individuals with disabilities was discussed, emphasizing benefits and risks associated with AI tools like chatbots. The text highlighted the potential for AI to revolutionize healthcare, accessibility, communication, and advocacy efforts for individuals with disabilities.Gary Marcus criticized the release of GPT 4.5 as underwhelming and falling short of expectations. He argued against pure scaling in AI, suggesting that alternative strategies are needed for significant improvements in AI development.These developments highlight the dynamic landscape of AI technology and its implications across various sectors. Stay tuned for more updates next week.

Thanks for listening! Follow us on Twitter, Instagram and Linkedin