Oscar Carrera Maya’s Post

Commenting to give visibility. Alejandro is a great professional and a team player.

🌐 Tech Giants and the AI Market: Balancing Innovation and Regulation In the ever-evolving landscape of artificial intelligence (AI), tech giants are vying for dominance. However, their strategies sometimes toe the line of antitrust regulations. This week, The Economist’s Schumpeter column highlights two key areas where regulators “trustbusters” are keeping a close watch: 1️⃣ Tying Businesses into Their Products: Some tech giants attempt to lock businesses into their ecosystems, making it challenging for competitors to thrive independently. NVIDIA, for instance, has faced scrutiny for tightly integrating its powerful #GPUs (Graphics Processing Units) with proprietary software. 2️⃣ Disguised Acquisitions through Generative AI Investments: Another tactic involves large investments in generative AI companies. Microsoft’s 49% stake in OpenAI and its absorption of Inflection AI’s staff exemplify this approach. Regulators are now accelerating investigations and sanctions, recognizing that even a small delay could lead to total market dominance, impacting consumers and enterprises. 🔍 The Capital-Intensive Quest for LLMs: Building Large Language Models (#LLMs), the backbone of generative AI, demands substantial resources��computing power, data, and expertise. These investments are accessible primarily to tech giants, prompting them to safeguard their interests by holding exclusive stakes in gen-AI startups. 🚀 Nvidia’s Valuation Surge: Nvidia’s skyrocketing valuation, recently surpassing Microsoft, echoes the dot-com bubble era. As AI hype intensifies (to try to reach control in next Internet wave as the one Google has now through its early search engine dominance), valuations become complex to align with future cash flows. In this race, balancing innovation and regulation is crucial. Let’s keep an eye on how these giants navigate the AI frontier. 🌟 #AI #genAI #LLM #AntiTrust #Innovation

Great session yesterday evening by Aric Rosenbaum (Chief Technologist at Red Hat) on "AI in a Hybrid World" at the IBM Watson premises in Astor Place, NYC. We had a valuable discussion about: 💪How smaller models, tailored to specific datasets or use cases, can outperform large language models (LLMs) and the importance of an architecture capable of training different models. This approach can significantly reduce costs. 🏦 The concerns banks have when using OpenAI, particularly regarding privacy and the use of data to train LLMs. 📈The importance of developing metrics relevant to the use case to quantify the quality of AI models. Thank you, Juan Jose Floristan for the invitation! PS: As I was leaving, a strong thunderstorm crossed the city, resulting in these beautiful pictures. #AI #HybridCloud #LLM #DataPrivacy #DataGovernance

After a break from posting about recent AI news, here’s my update for the week. I found three intriguing articles in The Economist, all in the 'Business' section, not in 'Science and Technology' one. 1️⃣ Retention is all you need: This article discusses the fierce competition for AI talent driving R&D efforts. OpenAI has lost half its workforce (about 100 people), while Google and Microsoft are aggressively hiring, with Microsoft recruiting most of Inflection AI’s staff. Interestingly, there's a trend of talent moving from tech giants (Google, Amazon Web Services (AWS), Apple, Meta, and Microsoft) to smaller players or rising stars like NVIDIA. 2️⃣ The LLM summer sale: This piece explores how hundreds of Chinese companies are developing Large Language Models (#LLMs) for their ChatGPT-like chatbots. Unlike OpenAI’s leadership in the West, the competition in China is driving prices down. Chinese companies offer LLMs at 1% of the cost of GPT-4 Turbo (pricing trigger 1,000 tokens), aiming to gather more data for training their models. They use less efficient chips due to American sanctions limiting access to advanced AI chips. Huge consolidation is expected. 3️⃣ Schumpeter Column - Tim Cook's prayer for the almighty: This article delves into Tim Cook’s generative AI strategy for Apple. Apple's approach focuses on integrating their own model with new M-series chips, using #ChatGPT only when necessary. This strategy could potentially drive more frequent iPhone upgrades. Will I need to change my cell phone soon? Stay informed on the latest AI developments! #AI #Business #Technology #IndustryInsights #ArtificialIntelligence #TechNews #TheEconomist #Apple #Google #Microsoft #OpenAI #Nvidia #ChinaAI #Innovation

Claudio, for the time being it isn’t neither intelligent nor generative, but it is already now a great ‘copilot’ to complement, help and boost human creativity. Indeed the next step, close to what some people label as transhumanism, may be indeed scary. #AI #genAI #copilot #transhumanism

Very valid questions, Claudio. Neutral hosts have been tested recently in other countries, not very successfully, but conditions are changing. And hyperscalers are starting to show less appetite for telecom business, as per Microsoft announcement yesterday to reduce its workforce in ‘Azure for Operators business’. #AI is their priority, as returns are higher and seem to arrive quicker Shaky times ahead of us!