Lugh

It's fascinating how similar to the brain's structure this is and how it also mimics its functions. Current AI models built on silicon chips are often assumed to lead to human-level AI - maybe that might be more likely to come from something like this? It also makes me wonder if this might be the right approach for interfacing with the brain for brain implants.

Africa is the last big part of the world to be electrified. It's interesting that it will be doing it differently, and that the solution seems to be shifting decisively to decentralized renewable energy.

Africa, like many other places, is plagued by corruption. I wonder if the decentralized aspect of mini-grids is a counter to that. It's harder for the people at the top to skim off their cut when mini-grids are so small and being built at the local community level.

The combustion of methanol produce less C02 than natural gas.

The idea here is that methanol would be stored for times the renewable grid wasn't producing enough electricity.

How often would that happen? It would vary a lot depending on other factors. Australia has enough sun to never need an option like this. How often would Germany use it? Depends again. How much other grid storage options might they have - pumped hydro, lithium batteries, etc, etc

Overall, they calculated that the cost of electricity from the grids would be between €77 and €94 per megawatt-hour delivered. That is well within the range that grid operators pay today to balance supply and demand via natural gas-fired power plants.

The other interesting thing about this idea is that it requires no new technology to be developed. It will merely need existing infrastructure and technology that produces hydrogen to be modified.

What's interesting about this is that it works alongside the other major solar tech - PV (Photovoltaic cells). PV can be used to provide power in daylight hours, and CST can be used to store it for night.

Also interesting to note how much shorter the ROI for investors will be . This is a key factor in what gets built and doesn't, and the chief reason new nuclear is rarely built anymore.

Flesh and blood guide dogs are expensive and hard to train, and thus only available to a small minority of the visually impaired. Quadruped robots are an obvious alternative. It often gets lost in the wider debate, but disabled people of all kinds will probably get more benefits than most from AI & robotics.

This is a fascinating read. It was interesting to hear him say that all of the current problems with factuality and hallucinations are solvable, and he sees the route to doing that.

He was less convincing in discussing how to constrain the power of an AGI that is smarter than us. His solution is to make sure it understands ethics. That idea has lots of weaknesses, and the guy interviewing pressed him on that a few times, and he seemed to fudge his responses.

It's interesting to look at people's past predictions. He said he predicted the 2028 date for AGI back in 2008. Not only was he able to say it hasn't changed, but he was also able to point out everything he expected has gone to schedule in the 2008-2023 timeframe. It makes that 2028 prediction more credible.

I'm curious what the route to market and commercialization is with this. That always seems another set of challenges once breakthroughs like this are made in the laboratory.

• it will probably end in violence.*

I have my doubts about that. The recent covid pandemic showed us how quickly people can adapt to radical change. How the civilian populations in Europe dealt with Word War One & Two provides further examples.

People are just as likely to adapt surprisingly quickly ; "Keep Calm, and Carry on" the well know poster was from 1939 in Britain at the start of WW2.

One of the age-old problems with markets is their tendency to renterism and monopolies if left unchecked. Now we have an added problem. Software, given ever-accelerating power by AI, may come to accelerate the capture of more and more of the economy to these forces.

This legal action is an attempt to address that, but is it enough? A recent economics paper points out that although renterism ( defined as the capture of scarce assets by groups who do nothing productive with them, and merely extract rents) is growing in Western economies, little is being done about it.

We seem to be heading for an economic reckoning with AI & robots and work and human jobs, maybe this issue will be dealt with then.

Interesting paper, but I doubt its central proposition has much likelihood of success. We're headed for super-intelligent AGI whether some people like it or not.

For a country that we often hear wants to win at dominating AI, this seems like a classic case of shooting yourself in the foot.