MIT Technology Review 3/2024: Superintelligence through AI models within reach?
When OpenAI boss Sam Altman was briefly thrown out of office in November 2023, there was an explosive report in connection with the personnel. The company has apparently achieved a breakthrough in general artificial intelligence (breaking latest news) with the “Project Q*” system.
Advertisement
This cannot be verified, but it shows that it is time to deal with the possible consequences: What will it mean if we one day have to deal with a human-like artificial intelligence? The models still work on the basis of pure statistics. They don’t understand what they are generating. But could a new form of intelligence develop from the complexity of the models?
Assessments from experts
TR editor Gregor Honsel asked around and asked those familiar with the scene around the world for their assessment – and the answers are as diverse as the personalities, including Katharina Zweig, head of the Algorithm Accountability Lab at TU Kaiserslautern, science fiction author Andreas Eschbach and Sebastian Thrun, ex-Google Vice President. While some believe a breakthrough will happen soon, others believe it could take decades or never happen.
Inventory of artificial intelligence
When even experts disagree, the question arises as to what constitutes breaking latest news. Can we even measure “real” intelligence? What must an breaking latest news meet to be considered such? TR editor Wolfgang Stieler set out to find answers.
AI in math lessons
Mathematics, among other things, is a field that researchers use to evaluate progress in AI development. This is not about the mathematics that goes into the systems’ algorithms, but rather about the question of whether AI models are capable of solving mathematical problems. Our author Eike Kühl describes in his text how good they are currently and what the challenges are.
Impact of AI in social
Rapid advances in AI are raising hopes for machines that can be friends or even lovers. In Japan you have been able to see what something like this looks like for a long time. Our author Martin Kölling provides the insight.
After the guesswork surrounding OpenAI’s possible breakthrough in the field of general artificial intelligence, it’s time to consider the possible consequences: What will it mean if we one day have to deal with a human-like artificial intelligence? Highlights from the magazine:
Further highlights of the issue
Tiefsee: How an adventurous math teacher, an enterprising physiologist, and a vast network of underwater caves in Florida led to a technological breakthrough in deep-sea diving.
Read MIT Technology Review here:
(lca)
To home page