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.
The per-kWh capacity cost at the link for hydrogen is very high compared to others I've seen. I wonder at the assumptions going into it. Are they assuming above-surface compressed hydrogen tanks, or liquid hydrogen?
Ultra cheap thermal storage promises cost at least an order of magnitude below that.
> 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.)
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.
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 is exactly what I’ve been thinking as I see LLMs being applied to all these complex problem domains. Humans did not conquer the world because our intelligence can solve every problem, we did it by using our intelligence to (1) break down complex problems into small, manageable pieces and (2) designing tools and machines that were exceptionally good at efficiently solving those subproblems.
The other recent example that comes to mind is the paper that explored the reasoning process used by LLMs to answer trivia questions like “Name a national capital whose letters can be rearranged to spell a common greeting in the language of a neighboring country.” (answer is Hanoi by the way)
The LLM responses show that they intuitively grasp the algorithm for answering such a question, but then they basically run the algorithm in their own thoughts (self-talk) which is horrendously inefficient.
Put differently, natural language reasoning is brilliant at turning the messiness of the real world into well-defined abstractions, but as soon as that is done it needs to hand off the task to a machine. For “solved” problems this might be a formally specified machine, but it could also be another class of model such as AlphaZero (along with a proper specification of the problem the “subcontractor” is to handle).
Can you elaborate a bit here? Neglecting _detailed_ info about storage conversion costs, etc, it's tough to understand the "all-in" cost for storage over time.
At home, I'm fanatical about using bog-standard AA/AAA rechargeable batteries for as many things as possible (anything with a micro-USB charger is basically "e-waste waiting to happen"), and thinking through any kind of home-supplement for solar, batteries, etc. makes me think that the "waste" of house-scale / grid-scale batteries for storage makes the math not work out.
Rough googling puts ~30kWh batteries at ~$15-30k, which: even if you think of it as having a 30-year service life, still works out to ~$50-100/mo in just battery depreciation.
Similarly with cars (eg: PHEV). First 5 years? Great! Next 10 years? ...a ticking time bomb of "must be replaced" with the battery representing an exorbitant percentage of the vehicle value. $500 of tires on a $5000 car is one thing, but a $5000 battery on a car seems like a net negative environmentally and financially?
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.
Your argument mixes "easy" and "cheap". Storage can be "easy" and you have plenty of choice: batteries, pumped, thermal, chemical, name it. If you have a large enough east-west country (hint), you might also invest in transport infrastructure to move energy where it is required so you don't have to store as much. This is also "easy", maybe moreso than storage because we've been doing it for so long. As to whether these easy things are cost effective when compared to other solutions is a completely different issue.
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.
Being expensive is the _easiest_ thing for most governments. It's often a requirement! i.e. As a contractor, bid too cheaply and you will not be taken seriously.
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).
Depending on your energy mix, you can go much further than 20% solar before needing storage. What's actually necessary is not storage, but generation which can ramp down and ramp up fast enough to compensate for the predictable daily ramps of solar generation (on top of the daily power use ramps). AFAIK, hydroelectric is one of the fastest (so much that pumped hydroelectric power plants can be used as storage), while coal and nuclear are among the slowest.
For cooling (datacenters and other) you can chill water when the sun is out. And mostly passively cool the heated water when the sun is down if water is scarce. Plumbing and insulation for cooling reservoirs probably degrades a lot slower than batteries. I'm not sure about chiller plant wear and tear vs battery degradation
In terms of logistics, you wouldn’t need to raise genetically modified pigs. You could potentially scale production on demand and it doesn’t require the environmental footprint of livestock.
It looks like it has dropped 4% over the last week, so while the monthly trend is upwards, I'm not sure this launch, or the scrub a few days ago, have improved things.
[1] https://www.authorea.com/users/960972/articles/1329770/maste...