2,852 karma · joined August 8, 2013
http://explicable.ai
benoit.paris.753@gmail.com
https://www.linkedin.com/in/benoitparis/
Kismet: 3b638cc74046887b5af69269b5c351d53acf4b702871f1c13691369db2917550
I'll take a try from the geographical determinism viewpoint, which I believe in 100%:
* Have natural resources in vast quantities: temperate arable land (cold enough to have winter 'resetting the fields', hot enough for high yields), oil. Cheap oil means cheap mining, cheap cost of transformation, cheap transports. Cheap everything.
* Unify the country from natural border to natural border, yielding very little need for domestic defense
* Be somewhat big, so that you can bootstrap any innovative company to a massive market and economies of scale quickly. Big also yields resiliency
* Give relative freedom to compete, and give it early in history
These 4 item are IMHO the base of many others which reinforce this strength (WW2 winner with no destructions at home, petrodollar enforcement, Hollywood machine, tech first-mover advantage)
----
Comparatively:
* Middle-East, Africa, India, Central America, SE Asia, don't have the 'winter reset'. Nor do they a need to develop a culture that had to plan and organize for a whole year; one part being food-rich, the other quite not
* Russia has huge unsafe borders (mandatory defense spending), elongated country (need for very creepy/scary internal security services, freedoms are quite limited)
* Europe has depleted its natural resources (but benefited from the industrial revolution then!), was divided for a long time (no big market), has no oil
* China has little oil, had a population matching its food production with very little to spare, had low freedom to compete
* Other parts are not big (Oceania), nor were populated early (South America, Australia)
----
For the future, the main changes will be fossil fuels depletion, and climate change. Here is what could happen:
* De(petro)dollarisation, as oil depletes and is less used in the industry
* World going a tad multipolar. This means less pax americana, more wars
* Resource wars shifting from oil to key materials for electrification like copper, lithium, nickel, cobalt
* Efficiency going down, massive simplification of lifestyles
* Lots of countries going bankrupt like Lebanon. Can't feed without importing? Don't have industry that can export? Enjoy your worthless money
* Fertilizers going more expensive with gas depletion. Lower amounts of higher priced food. Extensive world food programs
* All this right when we need resources to adapt to climate change
If you want the fees, you need to be aware of potential transactions; and also verify that these are valid (at pain of seeing your block be seen as invalid). So you need to receive that information.
Also if you want to publish a block you found, you need to not only listen to information but also publish it to others as well. So yeah, you kind of need to run an up-to-date full node to get both reward and fee.
Now, this does not mean that your miners need to be full nodes. These only need the block to be mined (digest of previous block, header, transactions, your public address to giver the reward to) and then they can vary the nonce at the end to start mining. But this block information has to be up quite to date, otherwise you're mining aimlessly. I've read that while switching to a new block, miner setups send block data empty of transactions to be mined: so as not to waste the few tenths of seconds it takes to upload new block data to miners.
So you can run an operation with a single central full node that sends blocks to individual miner devices.
Highlights:
* Seamless integration of hand-crafted rules with ML-build rules
* Intuitive, visual data/signal explorer: http://explicable.ai
* Implementation in Spark/Scala of treeinterpreter, currently used in production
* Participation to the FICO-Google Explainable Machine Learning Challenge
Location: Paris, France
Remote: open to it
Willing to relocate: for the right job, yes
Technologies: SHAP, RuleFit, Random Forest, Word2Vec, PCA, t-SNE, LSH, Scikit-Learn, Spark, Flink, Weka, Databricks, BigQuery, Hive, Postgres, MySQL, Oracle, AWS, Linux, Maven, Git, Java, Scala, Python, CAML, Elm, Javascript, Typescript, React, Spring, Primefaces, d3.js, Three.js
Résumé/CV: https://www.linkedin.com/in/benoitparis/
Github: https://github.com/benoitparis/
Email: benoit.paris.753@gmail.com
Blog: https://benoit.paris
---------------------------
Machine learning engineer, specialized in Explainable AI / ML. 13 years professional coding experience, 24 years personal.
Recent highlights:
* Seamless integration of hand-crafted rules with ML-build rules
* Intuitive, visual data/signal explorer: http://explicable.ai
* Implementation in Spark/Scala of treeinterpreter, currently used in production
* Participation to the FICO-Google Explainable Machine Learning Challenge
---------------------------
Tech: SHAP, RuleFit, Random Forest, Word2Vec, PCA, t-SNE, LSH, Scikit-Learn, Spark, Flink, Weka, Databricks, BigQuery, Hive, Postgres, MySQL, Oracle, AWS, Linux, Maven, Git, Java, Scala, Python, CAML, Elm, Javascript, Typescript, React, Spring, Primefaces, d3.js, Three.js
Résumé/CV: https://www.linkedin.com/in/benoitparis/
Github: https://github.com/benoitparis/
Email: benoit.paris.753@gmail.com
Blog: https://benoit.paris
Some others folks in the EU have chosen others means with lesser or unknown EROEI. But hey, their sovereign choices, their consequences. It's not like France is preventing them from doing nuclear.
I guess we'll see in a few years; but CfDs going over decades and being backed by governments I think it's safe to plan around this price.
About two hours is the cumulative time one must cater to the Dockerfile for a 3 weeks project.
But it requires institution insisting on reproducibility, and fostering best practices to make it even easier for the researchers to be compliant.
I get it that reproducibility can be quite hard for biology. But ML cannot be taken as an example of a hard problem.
Some very trivial Google searches will tell you their result is not out of the ordinary.
The size of its hype matches the size of the ICEs entrenchment in its moat. These have outsize influence on our economy, but climate change (and oil depletion) is quite inevitable. It takes irrational market cap to unseat that part of the economy being prisoner of its rent. And some allocators of capital have understood that.
I know this is not how us humans craft our thoughts, but maybe an AI can optimize to death the conjunction of these tools. The LLM just being an universal API to the core of formal logic.
https://prod-ec-us-west-2.video.pscp.tv/Transcoding/v1/hls/O...
----
Much better quality, slightly better latency. Seems like it comes from the Periscope transcoding infrastructure. Found it on reddit here: https://old.reddit.com/r/spacex/comments/17uyblj/rspacex_int...
But it is marked as dupe (HN seems to have already queried it and classified as broken link)
* 5 AM Pacific Time
* 7 AM Central Time (local)
* 8 AM Eastern Time
* 2 PM Central European Time
Live Media
* Official YT channel: https://www.youtube.com/@SpaceX
* Labpadre: https://www.youtube.com/watch?v=DIJTeeZj7k4
* NASASpaceflight: https://www.youtube.com/watch?v=mhJRzQsLZGg
* EverydayAstronaut: https://www.youtube.com/watch?v=6na40SqzYnU
* TheLaunchPad: https://www.youtube.com/watch?v=b0preOnsuo4
* TechniquesSpatiales(FR): https://www.youtube.com/watch?v=qI4cKoSD-Jg
Ressources
* Mission sheet: https://www.spacex.com/launches/mission/?missionId=starship-...
* Intro video: https://www.youtube.com/watch?v=18pCXD709TI
* https://old.reddit.com/r/spacex/comments/17uyblj/rspacex_int...
* 4 hours from now
* 5 AM Pacific Time
* 7 AM Central Time (local)
* 8 AM Eastern Time
* 2 PM Central European Time
Live Media
* Official yet channel: https://www.youtube.com/@SpaceX
* Labpadre: https://www.youtube.com/watch?v=DIJTeeZj7k4
* NASASpaceflight: https://www.youtube.com/watch?v=mhJRzQsLZGg
* EverydayAstronaut: https://www.youtube.com/watch?v=6na40SqzYnU
* TheLaunchPad: https://www.youtube.com/watch?v=b0preOnsuo4
* TechniquesSpatiales(FR): https://www.youtube.com/watch?v=qI4cKoSD-Jg
Ressources
* Mission sheet: https://www.spacex.com/launches/mission/?missionId=starship-...
* Intro video: https://www.youtube.com/watch?v=18pCXD709TI
* https://old.reddit.com/r/spacex/comments/17uyblj/rspacex_int...
Distances are not preserved though.
> Companies have said that the cost of building wind farms has soared because of rising inflation and interest rates, while the maximum price they can charge for the electricity they generate has been relatively low.
There's a very subtle link between all of these, which perfectly exemplifies the need to think holistically/systemically. Let's break these statements down:
1. "rising inflation and interest rates"
One might think that FED and ECB rule these, when they only follow the pace of the use of money. So what is money? It is a device to attribute value to products and services. So what is a product? It is matter that has been transported in your hands, and arranged in a very specific fashion, out of very specific materials. You know what is the only thing in the universe that can transport and arrange matter? Energy. The materials collected for a product are also energy, but not in the einstein sense, in the sense that if you had more energy you would be mining more of them. Nature gives us matter for free. It just lies there [1].
Money is created to try to make a 1:1 for energy. And as oil depletion is looming, as Russia attacked Ukraine, there's less available energy.
We have: -energy -> +inflation
2. "cost of building wind farms has soared because of rising inflation and interest rates"
Inflation is the fact that in money terms (also energy terms), products and services cost more. What's interesting is when you want to build energy producing things.
So for wind farms, we have: +inflation -> -energy
Now electrical energy is not the same as oil & gas energy. Except that it is. These become fungible on the electricity market, as gas can produce electricity. And using gas for electricity means not using gas for other things. Oil and gas prices are linked (being fungible for heat, and in chemical plants). So that means less oil, which is used in windfarm production heavily (transport of copper, transport and installation of the blades, etc.).
So we have a loop: -energy -> +inflation -> -energy. This loop is a differential equation, and it has an exponential solution. It is exponentially decreasing or increasing? And also let's consider that we might want energy to do other things than building device to produce energy. Doing energy for the only sake of doing energy is not desirable, and some [2] call it the "Mordor economy".
Which bring us to the EROI: Energy Return Over (Energy) Invested. That's what rules this exponential to increase or decrease.
3. "while the maximum price they can charge for the electricity they generate has been relatively low"
Well isn't that talking about EROI? Meaning that market prices are not enough to sustain the differential equation in an increasing manner? The increase in price they got from the UK government just means one thing: the part of the economy dedicated to use energy to produce device producing energy must increase. They need it so that the wind loop is not exponentially decreasing.
This also means the share for other uses for energy must decrease. This means less health services, fewer products, or lower quality ones. Less heating, less transport. But the wind energy will be sustained.
----
And this will get waaay worse:
* There's gonna be less oil, less gas over the years
* Even the gas-producing UK is saying they had to make more room for the wind farm sector
* We'll want to electrify a lot more, which means even more share for the energy sector that might have EROI problems
* That's without energy storage, which wind and solar need
---
TLDR: This differential system is what, IMHO, is why we desperately need to change our minds towards systems that have a better EROI. Nuclear energy has a great EROI.
[1] Energy is more slippery. It wants to spread out (that's called Entropy). You have to collect it and spend exponentially more (energy) the higher it is spread. For example the background radiation at 2.7 Kelvin will _never_ be useful.
[2] I can't recommend enough the works of Nate Hagens and Jean-Marc Jancovici. Here is a primer: https://www.youtube.com/watch?app=desktop&v=WYLKtz5A5pA
---------------------------
Machine learning engineer, specialized in Explainable AI / ML. 13 years professional coding experience, 24 years personal.
Recent highlights:
* Seamless integration of hand-crafted rules with ML-build rules
* Intuitive, visual data/signal explorer: http://explicable.ai
* Implementation in Spark/Scala of treeinterpreter, currently used in production
* Participation to the FICO-Google Explainable Machine Learning Challenge
---------------------------
Tech: SHAP, RuleFit, Random Forest, Word2Vec, PCA, t-SNE, LSH, Scikit-Learn, Spark, Flink, Weka, Databricks, BigQuery, Hive, Postgres, MySQL, Oracle, AWS, Linux, Maven, Git, Java, Scala, Python, CAML, Elm, Javascript, Typescript, React, Spring, Primefaces, d3.js, Three.js
Résumé/CV: https://www.linkedin.com/in/benoitparis/
Github: https://github.com/benoitparis/
Email: benoit.paris.753@gmail.com
Blog: https://benoit.paris
Highlights:
* Seamless integration of hand-crafted rules with ML-build rules
* Intuitive, visual data/signal explorer: http://explicable.ai
* Implementation in Spark/Scala of treeinterpreter, currently used in production
* Participation to the FICO-Google Explainable Machine Learning Challenge
Location: Paris, France
Remote: open to it
Willing to relocate: for the right job, yes
Technologies: SHAP, RuleFit, Random Forest, Word2Vec, PCA, t-SNE, LSH, Scikit-Learn, Spark, Flink, Weka, Databricks, BigQuery, Hive, Postgres, MySQL, Oracle, AWS, Linux, Maven, Git, Java, Scala, Python, CAML, Elm, Javascript, Typescript, React, Spring, Primefaces, d3.js, Three.js
Résumé/CV: https://www.linkedin.com/in/benoitparis/
Github: https://github.com/benoitparis/
Email: benoit.paris.753@gmail.com
Blog: https://benoit.paris
dF = 0
d * F = J
∇F = J
[1] https://en.wikipedia.org/wiki/Mathematical_descriptions_of_t...
Join us at:
https://www.youtube.com/watch?v=60z_hpEAtD8
https://enkimute.github.io/ganja.js/examples/coffeeshop.html...