Futurology

AheadForm was formed in 2024, and this is at the prototype stage, not yet ready for commercial deployment. Still, it shows the direction of travel. Somewhat realistically human-like humanoid robots may not be too far away.

If many people are already happy with faceless AI "companions," imagine how much they might like this?

VIDEO - Face Robot: AheadForm Origin M1

One of the things that sometimes goes unappreciated about renewables is that they are a technology, not just an energy source. As such, they are subject to the same improvements humans make with technology. Coal and oil long ago reached maximum energy extraction efficiency, and any gains now are minimal.

Hydrogel.

Solar panels lose efficiency as they heat up. The new gel absorbs water from the air at night, and cools by "sweating" it during the day. In tests, this has given a 12% relative boost in power conversion efficiency. The gel may even extend the panel's lifetime. However, there are questions about how this gel will be used over the 20-30-year lifetime of a panel.

Double-Sided Panels.

Tongwei has achieved a record 91.7% bifaciality in their solar panels, meaning the back of their panels is 91.7% as efficient as the front. This is significant because they've done it with cheaper technology that was supposed to be inferior. Most installations see about 10-20% more power from bifacial panels. The exact amount depends heavily on the setup.

Hydrogels keep solar panels cool, efficient, and durable

Tongwei achieves 91.7% bifaciality factor for 722 W TOPCon solar module

I wonder how much of this is down to AI? Maybe not much yet. Concerns about it and employment have only started going mainstream in the 2020s. That suggests there is more decline ahead for people's regard for the worth of college education.

It's striking how much opinions differ according to politics. 39% of Republicans rate college as "Not too important", versus 9% of Democrats who feel the same way. The article wonders if the perceived left-wing bias of colleges is to blame. But if right-wing people desert colleges, won't that just make them more left-wing? The student body certainly will be, and that's where the future staff members come from.

Perceived Importance of College Hits New Low: The percentage of Americans saying college is "very important" has fallen to 35%

This is interesting, but I don't know if it's all that significant. The swing towards right-wing authoritarianism makes a lot of the political questions very predictable to answer. Some relating to weather events, I would expect AI to be best at, as they're data crunching exercises.

Metaculus Cup Summer 2025

British AI startup beats humans in international forecasting competition: ManticAI ranked eighth in the Metaculus Cup, leaving some believing bots’ prediction skills could soon overtake experts

James Reed, chief executive of Reed, told Times Radio that his site advertised around 180,000 graduate jobs three or four years ago, and this is now down to 55,000.

He encouraged aspiring families to encourage their children to look into manual labour jobs as AI increasingly automates aspects of white-collar roles.

"The direction of travel is what worries me. Some people might say, well, that’s your business. But every other business is saying the same thing, that far fewer graduate opportunities are available to young people,” he said.

But guess what's a few years away? Cheap humanoid robots powered by AI. So even the manual labor jobs will start shrinking. Approx 750,000 people in Britain have jobs that are primarily driving vehicles; self-driving vehicles mean their days are numbered, too.

What we aren't seeing yet is these facts seriously impacting politics. When will that happen?

Graduates face ‘white-collar’ recession in jobs market

"Pakistan, which has for years treated gas generation as the backbone of its power network, has been asking suppliers to defer shipments of liquefied natural gas after a surge of solar imports suppressed grid demand. Saudi Arabia is facing one of the fastest declines in petroleum usage anywhere as photovoltaic farms replace fuel oil generators."

Analysts are talking about a supply glut of oil for 2025/26 lowering oil prices. Are we finally at the point oil use is going to start declining? Fingers crossed, let's hope so.

Meanwhile, China is almost single-handedly building the world's replacement.

China’s Marshall Plan is running on batteries: Beijing’s green energy projects are bringing jobs, growth and cheap electricity to the developing world

Social media algorithms are designed to make you angry, and the old media is only interested in sensation or 'if it bleeds, it leads.' So you might be surprised to find there's lots of good news in the world.

Here's some - globally, more and more land is being rewilded and going back to nature, and the trend looks like it's permanent. Decades-long productivity trends mean more and more food is being produced per square kilometer. With lab-grown meat and vertical farming in our future, these rewilding trends might even accelerate. Even if the human population finally peaks at 9 billion or so in a few decades, it won't reverse the trend.

The rewilding milestone Earth has already passed

This is a novel approach to geothermal energy. Existing methods, like industry leader Fervo, rely on drilling into hot rock and interacting directly with underground reservoirs. Rodatherm's method is entirely closed loop. It says that its closed-loop, refrigerant-based approach is 50% more efficient than a typical water-based system. However, it may have higher costs with drilling and installation. It just received $38 million in funding to build a pilot project in Utah to begin testing out its tech.

According to Wood Mackenzie, Fervo’s Cape Station project is demonstrating commercial viability at around US $79 per MWh without subsidies. Though it's working toward a target of US$45/MWh by 2035.

Rodatherm Energy wants to make geothermal more efficient, but will it be cheaper?

The video linked below is very interesting. In particular, look at how quickly the robot rights itself at the 6-7 second point after falling over. We're used to humanoid robots being slow and cumbersome, but no human can match that speed and agility.

That's Unitree's G1 robot. The developer version costs $40k, but the retail version is $16K, and they have a simpler R1 model for $6,000. The 2030s are likely to be filled with millions, and then tens of millions of these, many costing less than $10k. They will be far more affordable than cars, and far in advance of what we see in this video.

Video - "Violence tests" Professor He Kong's team from the Active Intelligent Systems SUSTech ACT Lab

The world’s biggest maker of sensors for self-driving cars has poured cold water on the chance of rapid growth for fully autonomous vehicles, saying society and regulators are not ready to accept deaths caused by machines that drive themselves.

“Close to one million people lose their lives every year to car accidents. If a technology company builds a vehicle that kills one person every year, that’s one-millionth of the difference, but it will have trouble to survive,” said Li in an interview."

I suspect the biggest obstacle to fully autonomous vehicles is the backlash against the unemployment they will cause. Safety will be used as an excuse to bolster that narrative. My guess is that by the 2030s, it will be clear to most people that they are far safer. They already are now, and they will be far more advanced then.

Top sensor maker Hesai warns world not ready for driverless cars

There are lots of reasons to think Chinese AI will come to dominate globally, and the Chinese government's latest directives on AI seem to make that more likely.

First, there's no mention of AGI or superintelligence. The only other nation's AI likely to dominate is the US's. China's approach to AI is profoundly different. Where the US AI leaders are focused on reaching AGI first, the Chinese are focused on the widespread integration of today's AI across all levels of society and their economy.

That means China can't help but export its AI standards. They are the world's manufacturing hub. This AI approach will be built into all their exports and thus spread around the world. EVs, robots, electronics, renewable energy infrastructure, etc, etc - all will have Chinese AI.

The Chinese make most of their AI open-source and free. They are more focused on the money they can make on top of that. Google with the Android OS is a good analogy. This will encourage global dispersal, too.

Finally, the Chinese have the advantage of having detailed plans and the ability to stick to them and implement them. Many Westerners favor as close as they can get to complete deregulation and the absence of any plans. The disadvantage of that approach in the 21st century, is that the Chinese and their planned joined-up thinking tend to leave you behind in the dust, while they get ahead and get things done.

The AI Plus initiative – China’s blueprint for AI diffusion

Data centers aren't the sole cause of Ireland's high electricity prices, but they do contribute to them. The biggest cause is Ireland's reliance on imported natural gas.

That said, data centers are heading for 30% of the country's electricity use, and they contribute significantly to high prices. Effectively a subsidy from Irish consumers to Big Tech. There are other externalized costs, too. E.g. Supporting Big Tech data center infrastructure is delaying house building. Ireland is lucky in that most of Big Tech pays its European taxes to the Irish government, so there's a quid pro quo here. But that is less true for other parts of the world.

Some people think AI may need as big a share of other countries' electricity - who should be paying for this?

Government warned of rising household bills as data centres strain grid

Full details, including a 27 page PDF are at this link

The study tested whether genetically engineered senescence-resistant mesenchymal progenitor cells (SRCs) could slow or reverse aging in primates. By enhancing the activity of the longevity-associated gene FOXO3, the researchers created stem cells more resilient to stress and senescence.

This reversed ageing across a broad series of markers, including the brain, skin, bones, internal organs, and reproductive system.

The study used cynomolgus macaques aged 19–23 years, which they said is equivalent to 57–69 years in humans. I don't know if you can directly scale up the improvements to "human years", but if you could, it seems this would be the same as reversing human aging by about a decade for people in their 50s and 60s.

Senescence-resistant human mesenchymal progenitor cells counter aging in primates