91É«ÇéƬ

The article on using artificial intelligence to tackle scientific challenges paints too rosy a picture of its accomplishments up to now. AlphaFold sometimes inaccurately predicts protein folding. The matrix multiplication methods found by AI only achieve a 10 to 20 per cent improvement for certain small matrices and only when using certain hardware. The sorting …

I laughed when I read that AI could help create viable nuclear fusion power, with "energy so cheap that it could be given away for free". Nuclear fission was supposed to be "too cheap to meter" as well. In fact, it has always been the most expensive energy fed into the grid. The type of …

While reading your article on the use of psilocybin to treat anorexia, and reflecting on the growing number of reports on similar studies of the beneficial effects of controlled substances, I wondered if we are seeing a "generation drug" effect? Are those who lived through the normalisation of the use of recreational drugs in the …

I spent much of my career trying to discover drugs to treat obesity. The class of drug that my team and I found to work well in rodents was ineffective in people. However, others found them useful in the treatment of overactive bladder ( 15 July, p 32 ). How I wish that I had …

Cosmic rays produce helium-3 and beryllium-10 in asteroids ( 22 July, p 16 ).

I doubt I am the only one alarmed by the idea of prescribing weight-loss drugs to children, as discussed by Clare Wilson. It comes amid studies showing that a rise in consumption of ultra-processed food is linked with a decline in physical and mental health ( 15 July, p 32 ). Manufacturers create ultra-processed foods …

Having visited Bolivia on several occasions, I found it spectacular. Moreover, as a scientist, I well understand the consequences of destroying its forests. However, I still have to question the nature of the commentary in your recent picture story on this. How can people living in higher-income countries suggest that Bolivians don't have the right …

You write that "until we know what consciousness is, there is no solid way of testing for it" in an artificial intelligence. There is, in fact, a simple way of testing for it, provided the AI is incapable of lying about itself, has colour vision and hasn't been trained on the concept of "subjective experience" …

You state: "Perhaps the relevant question isn't whether or not AIs can become conscious, but why we would want them to be conscious." If you see conscious awareness as the ability to express feelings, thoughts and intentions, then the importance of consciousness in robots that are to be useful assistants, carers or interesting companions is …

The consequence – or economic threat if you are personally affected – of AI is that armies of white-collar professionals, such as architects, lawyers and accountants, may become permanently redundant. You ask an AI: "Show me some designs for an east-facing beach house with a veranda and water views that backs onto a bush reserve." …