Isinlor

Ukraine war is very serious.

After Poland managed to collapse communist regime in partially free elections in 1989 where all elected sits went to democratic opposition the whole Soviet Union collapsed. Ukraine declared independence in 1991, all regions of Ukraine decided to stay with Ukraine in national referendum.

In 1991 Polish and Ukraine GDP were about equal, both countries have around 40 million citizens.

Poland immediately turned towards West. We rapidly liberated our economy - Balcerowicz reforms. We aimed at joining NATO and EU as guarantees of our independence from Soviets/Russia. After invading Poland in 1939 with Germans, last Russian soldiers left Poland in 1993. We joined NATO in 1999 and EU in 2004.

Ukraine on the other hand was divided between West and East.

In 2021 Polish economy was 3 times bigger than Ukrainian. More than million Ukrainians were economic migrants in Poland.

We made the right choice with turning immediately towards West with full force. Ukraine stumbled.

But they saw that Poland is doing so much better than they were doing. That's why they wanted to join EU. That's what started Euromajdan.

Now Russia is threatening all of us with nuclear holocaust. Every other month you can hear from the highest Russian leadership nuclear threats.

We have lost 20% of population in WWII started by Germany and Soviet Russia. Every fifth person killed.

We are taking Russian holocaust threats 100% seriously.

Yup, the models are smart, but are trained to follow standard human patterns for this type of questions. And even on hackernews vast majority will not think that they would need to correct for buoyancy when actually attempting the experiment in standard conditions.

They very often get popular "tricky" questions wrong because they saw it so many times that they switch from internal reasoning to memorization/retrieval.

When Can Transformers Count to n?

https://arxiv.org/html/2407.15160v2

The Expressive Power of Transformers with Chain of Thought

https://arxiv.org/html/2310.07923v5

Transformer needs to retrieve letters per each token while forced to keep internal representation still aligned in length with the base tokens (each token also has finite embedding, while made out of multiple letters), and then it needs to count the letters within misaligned representation.

Autoregressive mode completely alleviate the problem as it can align its internal representation with the letters and it can just keep explicit sequential count.

BTW - humans also can't count without resorting to sequential process.

Transformers are very bad at counting in one feed forward pass, you need to explicitly tell them to use a counter in autoregressive fashion like here:

https://chatgpt.com/share/6775cb37-4198-8007-82cb-e897220827...

Transformers are very bad at counting due to how their internals work. But if you ask them to use explicit counter the problem disappears:

https://chatgpt.com/share/6775c9a6-8cec-8007-b709-3431e7a2b2...

Basically one feed forward is not Turing complete, but autoregressive (feeding previous output back into itself) are Turing complete.

Reasoning is not a binary skill.

And failure modes of other types of reasoners do not need to be the same as the failure modes of humans.

It's widely accepted that reasoning is not a binary skill.

You can make mistakes and still reason. Very often people given the same premises will disagree in thier reasoning as we are doing right here.

BTW - the model may be wrong depending on the example. More voluminous objects displace more air and due to buoyancy are lighter for the same mass.

The proper way to ask it would be to ask which object has more mass.

The task of Russian navy was to invade Odessa and sabotage Ukrainian grain trade. They didn't even attempt Odessa desant, because they knew it was futaile. And they failed at sabotaging Ukrainian grain trade, but not for a lack of effort.

Otherwise these ships served as missile lunchers, but even then they had to widraw to ports far away from Ukraine, because of drone attacks.

Ukraine won the naval battle at tiny fraction of the Russian navy cost.

The ship has long sailed. EU decided that our laws will be applied maximally - to all EU citizens regardless of their location and for everyone within our territorial control.

And if other countries like USA have conflicting laws like Patriot Act then companies can be even forced to divest from EU or split their businesses so that USA branch can not enforce USA spy laws in EU. Google Privacy Shield.

Depends on the benchmarks. AI that can actually do end to end the job of software developers, theoretical computer scientists, mathematicians etc. would be significantly more impactful than this.

I want to see AI moving the state of the art of the world understanding - physics, mathematics etc. - the way it moved state of the art of the Go game understanding.

probably the atomic cooperation is even more important

That's why I don't understand why Budapest Memorandum signatories allow Russia for breaking the security guarantees.

Ukraine gave up fully the third biggest nuclear arsenal for security guarantees from Russia, USA, France and UK.

To me it seems pretty clear that if Ukraine falls then nuclear disarmament and nonproliferation is dead. And that my country, Poland, should seek to develop nuclear weapons.

On the basic premise EU is working wonderfully.

It's the longest period in our thousand years history that we Poles have not had to fight Germans.

We have lost maybe even 1/3 of population in wars with Sweden, 1/5 of population in war with Germany.

Poland didn't exist or 123 years under Russian, Prussian and Austrian occupation.

Given that as a baseline, I'm really grateful for the EU.

The Digital Markets Act is pretty damn well made regulation. Limited only to the very biggest companies, so called gatekeepers, and with wide executive rights given to EU Comission to prevent the act circumvention.

It really opens up a lot of walled gardens. I'm also looking forward to interoperability of chat services.

Pity that AI act is nowhere near as good. And if passed as proposed by EU parliament it will end up killing EU chances at competing on the emerging AI market.

Slavic languages do perfectly fine without articles. As Polish speaker I say leave them out ;) .

I played the Witcher 3 twice trough leaving a small bit of the second expansion unfinished just because I don't want this game to end.

I lost interest in Cyberpunk quite early. To me Cyberpunk is too American with overdone themes. Although the Cyberpunk anime is great. Simple story, but has really well put together characters and delivers emotional punch. I will have to try to get back to the game. Maybe on second try it will be better.

I like simulation-like open world games, but Witcher 3 was one of the very few games with a good story. To be honest I think the only one that I cared about. Especially the bloody baron questline was amazing storytelling. Heart of Stone was also great story based on Polish ballad Mr. Twardowski.

Cyberpunk is unfortunately nowhere near as good in terms of the story as the Witcher.

I can show you many countries with high fertility rates where people do not even have the idea of stable employment or social benefits.

Then there are rich countries, great education, great child care facilities, government support and low fertility rates.

Rich people are busy complaining about not being able to have children. Poor people just have them.

This is also well visible in post-communist countries.

During communism there was nothing in the shops, big families were living in small apartment with parents sleeping on beds opening from the wall. There was no disposable diapers. In general people had not much career opportunities.

And then communism collapsed, and as people became richer trough capitalism fertility rates plummeted.

And there was no policies to convince people against having big families.

The basic pattern is the same across the world poor countries have high fertility rates. Rich countries have low fertility rates. And as poor countries become rich their fertility rates go down.

No, I mean: https://en.wikipedia.org/wiki/European_Union_Emissions_Tradi...

It was introduced, then largely forgotten until prices started spiking.

Nobody will remember about it in few years. The same was with cap and trade system.

You are free to make your life as difficult as you want. You can move to Alaska and try to survive there from manual work.

There is a very good reason why people want to go from poor countries to rich and not the other way around. Poverty is not fun.

Dark matter is also believed to interact with ordinary matter in ways other than gravitational.

In realm of everyday experience neutrinos and proposed dark matter would be the same ghostly thing.

That's exactly what neutrinos are in terms of human experience. Ridiculous amounts of neutrinos are flowing trough the whole planet and trough you like we are nothing.

Tesla is supply limited, so they can put prices high. That's probably about it.

Humans make mistakes with rules too.

You are not explaining your internal processes (neurobiology, brain chemistry, neuron firing patterns etc.) that caused you to arrive at given moves. You will only provide proofs that your moves are good.

They are not explaining their internal processes (neurobiology, brain chemistry, neuron firing patterns etc.) that caused them to arrive at given moves. They just provide proofs that their moves were good.

Actually USA is using the metric system.

They just apply their weird conversions on top of the metric system, but fundamentally they use the definitions of the International System of Units.

E.g. Pound is legally exactly 0.45359237 kilograms where kilogram is defined by setting the Planck constant h exactly to 6.62607015×10^−34 J⋅s (J = kg⋅m2⋅s^−2), given the definitions of the metre and the second.

Previously kilogram was a block of metal stored in France. But the copies of kilogram were diverging and there was no way of knowing where the changes were coming from. So now the kilogram is defined with regard to the physical properties of the universe as opposed to specific object.

E.g. a meter is exactly the distance traveled by light in a vacuum in 1/299792458 seconds. Where second is exactly the duration of 9192631770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium-133 atom.

People overestimate what's possible short-term and under-estimate what's possible long term.

Current AI models have three main limitations:

- rapid skill acquisition. Example: someone invents a new programming language. Something like Elm or Rust. Humans can start using it right after reading a "quick start" page and few tutorials from the authors of the new language. How much training data will GPT-style models need to start using that language? A lot, like output of hundreds or thousands of people. This needs to go down by a factor of 10x to 100x to match humans.

- agency, or taking actions guided towards a goal. Example: Can you ask GPT-like model to book a flight ticket for you? Help it help you learning Photoshop? New IDE? Test your latest app or indie game and find bugs? No. The ability of current models is still not good enough to be useful to an average person with regard to interacting with outside world.

- acting in the physical world. Example: An average human can learn to drive enough to pass a driving test in the order of 100 hours during a driving course. How far away are we from a system that can control humanoid-style robot to learn to drive in the order of 100h? Currently, we can't do it with billions of dollars with specially designed hardware. Using humanoid robot to control a car is not even useful as a benchmark for the state of the art machine learning systems.

IMO the currently existing systems like ChatGPT or Stable Diffusion are worth all put together in the range of $10B to $100B in the next 3-5 years.

Future systems that will address all 3 limitations mentioned above may be literally the whole reachable universe changing if we decide to build self-replicating space probes (Von Neumann probes). We know that there are physical systems that are not very intelligent but capable of exponential growth like viruses or bacteria (humans too).

The main limitation of biological systems is their adaptability, especially to lack of water. If robots can build other robots avoiding bottlenecks of human intellectual and manual labor then the robots are limited only by resources and energy available. We have plenty of both on Earth and Solar system is full of it.

Also, I'm just talking about human-like abilities. All of that is possible without involving concepts like superintelligence. Bacteria are not very smart, but they can multiply exponentially.

The spread of possibilities is enormous.

All of that hinges on your timelines on the 3 above mentioned limitations. It's like predicting when atomic bomb will be possible. It may not happen for decades, or there may be right this very moment someone with an idea that will make it all possible.