electricships

We have plans for (millions) of HPDE sea floaters which should last for 10-20 years (structural strength is not a major issue iirc)

Is there a pathway to replace these with recycled plastic something?

would love for this to come fr9m OceanCleanup plastics somehow

Any idea about the Untether accelerator pricing ?

e.g. tsn200 40W 500TOPS tsn800 300W 2000TOPS

Similar products [1,2] are approx $1-$2/TOPS (this is theoretical 8bit int ops, so multiply by x2-x4)

[1] https://www.axelera.ai/ [2] https://hailo.ai/#discriminative-ai

After AlphaZero, people tried to train various combinatorial solvers

IIRC, they were unable to exceed Sota of SAT solvers (which admittedly are the results of decades of research)

maybe we just needed bigger networks, or transformers, or just more training compute

Are current compiler optimisations limited by algorithm or by compute?

would a x1000 compute cluster provide a meaningful performance boost ( of the generated binaries)?

how much higher is the capex for deep sea floating structure ?

(i guess stuff like o&m, grid connection and yield are really project dependent)

edit: solar duck just announced 150mw combined pv wind project

we need all of the above

deploy existingg tech, work on the remaining problems

if there are angels here for deep sea clean shipping, kindly DM me

pros: global moat, feasible tech, price parity at $80/barrel , massive upside

cons: very early, capex heavy, long timelines

yeah nuclear cargo ships seem fancyful

if there are angel investors here for solving deep sea clean shipping, please email me

pros: global moat, commodity components, price parity at $80/barrel , massive upside

cons: very early, capex heavy, long timelines

here is my good deed for the day:

modern AI is just vector multipication. any AI chip is just 10,000s of very simple cores which can do vector float operations and little else. this also entails clever trade offs of shared cache and internal bandwidth.

(as a thought experiment, consider a naive million by million matrix multipication. this will take a single cpu about 1 year! how do we reduce this to 1s ?)

the end

Egypt recently completed an extension to the Suez canal

Turkey is also building a canal from the black sea.

you can look up the costs. both of these countries are very regulation-light...

What is the state-of-the-art for battery-electric ships?

the power is probably MVDC from the large battery. And these are large xxMW motors.

So will we end up usong inverters anyway?