gglon

The claim hinges entirely on the narrow definition of "large-scale energy generation." For the UK, with its high seasonal energy demand and low winter solar output, the cost of generation is almost irrelevant next to the cost of firming that power for 24/7/365 availability. While the paper[1] shows solar PV and daily-cycle batteries are getting cheaper, it also shows seasonal storage solutions like hydrogen are still an order of magnitude too expensive and inefficient (huge capex for electrolyzers/storage + poor round-trip efficiency). So providing reliable, 24/7/365 baseload power from PV + storage in the UK is demonstrably not cheaper than gas or nuclear today.

[1] https://www.authorea.com/users/960972/articles/1329770/maste...

> At the time of writing, the highest-ranking non-purely-transformer-based model on the LM Arena is Jamba, which is a transformer–state space model hybrid, at rank 96.)

Tencent's hunyuan-turbos, another hybrid, is currently ranked at 22. https://arxiv.org/abs/2505.15431

From the paper, section 2.1: minimize_θ,φ,Δ ||P_φ(Δ, E_θ(x)) - sg(E_θ'(y))||_1

where

y - full video, x - masked video, E_θ(.) - learned encoder (semantic embedding), P_φ(.) - learned predictor, Δ - learned mask (which patches in a video where dropped), sg(.) - stop gradient to prevent change, gradient propagation in E_θ'(.), which in turn is an exponential moving average of E_θ(.) ie. θ'_new <- τ θ'_old + (1-τ) θ. So the loss is applied only to the predictions of the masked patches while the encoder of full video follows the learned one. This asymmetry in learning prevents collapse of the encoder to a trivial constant.

It seems like a model is not included; you need to set an API endpoint in a configuration https://tmuxai.dev/getting-started#environment-variables

I was thinking, to build a large, efficient factory autonomously, one could use LLM as a high level agent that is using specialized tools. The overall strategy would perhaps look like following:

1. create a (intermittent) goal for a resource production

2. create a factory graph with calculated number of machines and number of resources required to transport between them. This would be done by using linear programming (factorio calculator)

3. somehow map the resulting graph to a hardware description language. Such that each entity would be mapped to unique logic component. And each transport lane would be mapped to a unique wire (most difficult)

4. compile to 2d FPGA layout using all the VLSI algos like partitioning, routing (hdl compiler)

5. map the resulting plan back to a concrete factorio design

This doesn't apply to energy grid scale projects. Also that is why energy sector in most countries is privatized. Take Germany with total energy costs that would be like ~10% of the federal budget (about 40 billion euros). Making energy even more expensive would only lead to further deindustrialization and lower gdp making it politically non viable. Also, you can't just throw money and buy storage like with typical government contracts - it requires building entire new industries (like Germany's struggles with hydrogen infrastructure).

none of those storage solutions are easy for grid scale. For batteries you need whole factories, pumped requires permits and tone of cement, chemical is like hydrogen with low efficiency. All of those are expensive which for government is synonymous with difficult.

pumped hydroelectric is storage; right now it is like 99% of world storage capacity.

Solar PV operates 10-30% of the time (depending on location). Without storage, this naturally limits direct solar contribution to about 20% of energy demand. Going beyond requires expensive storage solutions.

Achieving the first 20% of solar in an energy mix is relatively easy. Beyond that costs increase 3-5 times per energy unit mainly due to storage.

How does it compare with xenotransfusion from genetically modified pigs?

https://en.wikipedia.org/wiki/Xenotransfusion#Ethical_argume...

I guess the extended shelf life and more compatibility might be some of the pros of synthetic blood.

Also magnet for the pile works well: magnet:?xt=urn:btih:0d366035664fdf51cfbe9f733953ba325776e667

Andrej Karpathy would agree - recent talk: https://youtu.be/bZQun8Y4L2A

According to https://www.metaculus.com/questions/14270/spacex-starship-or... its chances to reach orbit in 2023 have just risen from 85% to 90%. Therefore generally I would call the launch a relative success.

Yes; metaculus is reputation based. Though in the comment section you will find links to other markets like https://polymarket.com/event/will-erdogan-win-the-2023-turki... which is money based market.

https://www.metaculus.com/questions/15036/2023-presidential-... - prediction seems like 50/50 now

In my opinion the biggest near term danger is that AI will make people overly trust it. And they will start treating it as a truth oracle. Oracle that will be controlled by some small group of people. Namely, it will become a perfect tool for propaganda and control.

That is exactly why I think one post per week and lack of global space is a great solution for a startup; it massively simplifies spam and other hostile content early detection and removal.

As I understand it a one way to disincentivise centralization is that when your node loses access to the network you lose more money if at the same time many other nodes/stake do not have access to the network. So it would be profitable to have an independent server and internet access.

In Mathematica 12.3: Integrate[Log[1 - 2 a Cos[x] + a^2], {x, 0, Pi}, Assumptions -> Abs[a] >= 1 && a \[Element] Reals] gives the correct answer -\[Pi] Log[1/a^2]

Also see Casual Entropic Forces [1] by A. D. Wissner-Gross [2] and C. E. Freer

[1] https://www.alexwg.org/publications/PhysRevLett_110-168702.p...

[2] https://www.alexwg.org/

> Isn’t it nice (...) advance our understanding of the fundamental forces of nature?

The criticism is that the ROI = (the potential advance) / (invested money) is too small. As you pointed out it can be argued that the real goal is elsewhere: in the engineering / technology or science of complex systems. Creating a big machine that verifiably works brings many unexpected advances in many fields. And what is more important it is a symbol of human advancement which unites people and inspires them to pursue science. Since humans need a clear goal to invest much resources in a constructive endeavor. It may not be that important though that the ROI in fundamental physics is small, as long as people work hard together. But it can be argued that the whole process begins to resemble religion more than a scientific method; which can be problematic as religions have a bad reputation among scientific community.

It may be so. But an underground pump hydro is just boring.

Eduard Heindl's gravity storage: https://heindl-energy.com/

Potentially reaches a storage capacity between 1 and 10 GWh. Lifting a (huge) rock by pumping water underneath.

Interview: https://omegataupodcast.net/299-gravity-storage/

The discomfort is probably the one and only thing you should really trust.

Still it is harder to deal with sophisticated AND not sophisticated attackers than with "only" sophisticated ones.

That would be the whole gain: set of attackers would be filtered down to sophisticated ones. That would be advantageous for the network since the more centralized potential attackers are, the easier it is to deal with them.

Yes, but not proportionately since an access to capital would be an additional cost. Eg. Imagine 2 situations when you can earn $100 by investment:

1. Buy sth for $100 end sell for $200

2. Buy for $1mln end sell for $1 000 100

Option 1. seems much more accessible.

Yes, then the overhead(+ incentive) should be subsidized by Peers - normal users that just use the network eg. for transactions. As a tx fee or inflation.

Webpage is also centralized. The whole point in creating PoW is to create a task that would be more expensive on the network than on the market. ie some of it would be "wasteful" - used only to secure decentralized network (see my other post)