⤴⤵ Up Wing/Down Wing #18

A curated selection of pro-progress and anti-progress news items from the week that was

In case you missed it .. .

✨🤖 The Age of AGI: the upsides and challenges of superintelligence (Tuesday)

☀️ My chat (+transcript) with economist Noah Smith on technological progress (Thursday)

✨ AI (but not supersmart AGI) and the world of 2030 (Friday)

---

Up Wing Things

💰 Chipmaker Qualcomm has approached Intel about a potential acquisition. While no official offer has been made, the deal could face hurdles like regulatory scrutiny due to the companies' size (an acquisition would mean rigorous antitrust reviews) and national security concerns. Intel, struggling with management issues and losing ground in chip manufacturing and AI, has seen its market value fall to $93 billion, while Qualcomm’s has risen to $169 billion. Qualcomm, a leader in smartphone chips, is unlikely to want Intel's foundry business but could benefit from their chip design, software expertise, and PC sales channels. (NYT)

🔋 Better grid-scale battery options are on their way. The market for grid-scale batteries is expanding as alternatives to lithium-ion technology, such as zinc-bromine, sulfur-sodium, and carbon-based flow batteries, offer safety, environmental, and economic benefits. While traditional lithium batteries require rare materials that can be hard to get (coming from regions with geopolitical and environmental risks), the alternative options are designed with accessible and fully recyclable materials. Those alternative battery designs are usually much larger in size, but they provide large storage capacities and safety features which traditional batteries are lacking. These technologies target long-term for more sustainable and cost-effective energy. (Ars)

💡 AI outpaced human idea productivity for original research. An AI ideas generator outperformed 50 independent scientists in coming up with more original and exciting research ideas, according to a study posted on arXiv:

The researchers prompted their AI tool to find papers relevant to the seven research topics using Semantic Scholar, an AI-powered literature-search engine. On the basis of these papers, the researchers then prompted their AI agent to generate 4,000 ideas on each research topic and instructed it to rank the most original ones.

Human and AI-generated ideas were evaluated anonymously, with AI concepts scoring higher in novelty but slightly lower in feasibility. While impressive, only 200 of the 4,000 ideas generated were determined to be genuinely unique. The study, however, has limitations, including focusing on a single research area, quick idea generation by humans, among others. Nonetheless, the study lays the groundwork for further exploration of idea generation by LLMs in research contexts. (Nature)