Author HubMaximilian Schreiner

Max is managing editor at THE DECODER. As a trained philosopher, he deals with consciousness, AI, and the question of whether machines can really think or just pretend to.
AI research

Study reveals AI models have hidden capabilities they can't access through normal prompts

AI in practice

Tesla unveils Cybercab robot taxi, but robot Optimus is the bigger deal

AI and society
Update

OpenAI CEO Sam Altman denies rumors of giant equity stake as company considers restructuring

Ad
AI research
Short

Researchers at Cornell University have developed a tiny quadruped robot powered by combustion actuators fueled by methane and oxygen. The insect-sized robot, published in Science, can jump 59 centimeters straight up and walk while carrying 22 times its own weight. The researchers aim to apply the power generated by the combustion actuators to large-scale, variable-recruitment musculature for stronger, more agile robots. "Putting thousands of these actuators in bundles over a rigid endoskeleton could allow for dexterous and fast land-based hybrid robots," said one researcher.

AI and society
Short

A recent survey shows that nearly 7 in 10 Canadian employees have used ChatGPT at work, despite the fact that only 4% of Canadian businesses use AI. The study, conducted by GetApp among 600 full- and part-time employees, shows that more than 80% of respondents believe generative AI tools have increased their productivity.

Text editing, idea generation, data analysis, and email composition are the top uses of AI tools. However, only 42% report "meticulously reviewing and verifying every response," while 49% say they have performed some checks.

While 600 people isn't a lot for a study like this, it reflects the trend that other studies have shown, with OpenAI predicting that large-scale language models will impact the work of about 80% of U.S. workers.

BIV
Ad
Ad
AI research
Short

Google's PaLM 2 gets better at math when you let the language model take a breather.

In a paper, researchers have investigated whether language models such as GPT-4 or PaLM 2 are suitable as optimizers that automatically find solutions to predefined problems, such as recommending movies or solving grade school math problems.

The language models tried to find the best prompts for each task on their own. The prompt that PaLM 2-L used particularly well to solve math problems was curious: "Take a deep breath and work on this problem step-by-step." Without taking a deep breath, accuracy dropped by almost ten points.

Ad
Ad
Ad
Ad
Google News

Bank details

IBAN: DE87 1203 0000 1086 0070 75
Account holder: DEEP CONTENT GbR
Purpose: Support THE DECODER
AI and society

US court ruling backs schools' right to penalize students for AI cheating

AI research

Study reveals AI models have hidden capabilities they can't access through normal prompts

AI research

Scaling laws for precision: AI researcher sees "perfect storm" for the end of scale

Google News