dr1337

Makes me wonder how much better these systems would be at online learning as opposed to systems that require separate training and inference. This is especially relevant in a dynamic environment for applications such as robotics.

Yeah, I don't think it's mentioned in the paper however in the New Scientist podcast it was alluded to by the lead author that there is currently work being done to compare. I'd reason that they are looking at wall clock time since ML systems can train in no time if done in a batch-accelerated process.

That's fascinating. It reminds me of this clip from Adam Curtis's All Watched Over By Machines of Loving Grace where they got lots of random people to play Pong in a theatre in the 90s:

https://vimeo.com/78043173

So it looks in the paper they're using a theoretical framework called the Free Energy Principle developed by Prof Karl Friston at UCL which postulates that brains have evolved to minimise the amount of information surprise between their own internal generative model of the world and the observable world.

The paper uses this as the basis of learning and it looks like the neurons are given a random (electrical) stimulus when they miss the ball and a predictable stimulus when they hit it.

Yeah but there are lots of companies doing this already including Intel that has poured millions into their Loihi neuromorphic chips.

They're already doing it with mice! - https://www.pymnts.com/cryptocurrency/2021/crypto-trading-ro...

Found the visualizer they used for the videos here - https://spikestream.corticallabs.com

Very cool WebGL implementation. I hope they release the dataset so that we can visualize all the games and maybe do our own analysis on them.

The real MVP. Thanks!

Don't be surprised in the coming months to read 6 out of your 11 Onion headlines coming out as real scientific work.

Really? I haven't found it an issue to get friends or family to sign up for Element. In fact it's even better because you don't need to use your mobile number for that.

Some people can't fathom how people in China can live under perceived authoritarianism are also those who use iPhones and support Apple's decision to lock down the AppStore and iOS because of privacy and safety safeguards. See the parallel there?

Monopsony is also deemed to be anti-competitive and illegal. Technically as an app developer or service provider, when you're selling an app or service on iOS, you're not selling it directly to the user. You're actually selling it to Apple who then on-sells it to the user. Because of the way the walled-garden has been set-up, there's no alternative way of reaching customers unlike in Android where a customer can have the option of multiple app stores or the ability to just sideload the app.

Apple go down the Samsung route and license RDNA tech for their ARM GPUs - https://www.tomshardware.com/news/amd-rdna-exynos-samsung-so...

How did you get Looking Glass to work with macOS? I don't think it's supported as host OS.

Are you tuning hyperparameters when learning a task?

I guess if the stem-cells are taken from you as a donor then they would be genetically identical to you as much as twins are genetically identical individuals?

Ditto for Australia too where there was a hard swing towards the right.

It's funny how Fortnite which is owned by Epic which is owned by Tencent is turning out to be the 21st Century's new Opium Den.

The cost blowout in healthcare isn't due to diagnosis or care but administrative costs. All these diagnostic AIs won't significantly reduce the cost of healthcare nor improve quality. What we really need is automated administration and billings to really move the needle and that's more human bottlenecked than technology.

Exactly the reason why unbridled capitalism is terrible. I'd imagine that if we had a genocide going on today, a company like Bayer would be rewarded by its shareholders for manufacturing Zyklon B since the only reason for a corporation to exist is to generate profits right?

So much for Silicon Valley corporations and their "culture" and founding myths.

I guess people don't even remember that Google once you used to have an informal mantra of "Don't be evil."

It does but not where it matters. You want places that are willing to fork out big money for flagship phones. An example of this would be China vs India.

China is a better market to be in to maximise profitability as the Chinese are more likely to fork out USD 799 for a phone than the Indians. It's almost mind-boggling how one company can capture so much share of profits.

I've never really understood the US-centricity of most tech companies. Sure the US will always be the single biggest market but times are changing. With the rise of the middle class in the emerging economies and the global redistribution of wealth from the West to the rest, there are huge opportunities to leapfrog US-centric companies.

Apple is probably the best example of a company that realises the opportunity of building a supply chain that can satisfy the demand of their goods on a global level.

Fair enough - having said that, I'm pretty sure you would have done alright with or without Tinder.

The issue is selling hope to incels. That's where the real money's at.

Hope - the best commodity that you can sell.

It's no wonder Tinder is up there on the Play Store as the second highest grossing app alongside all the online gambling apps. That's because Tinder and all the online dating platforms are exactly designed to mimic gambling.

You can't win if you're not in the game.

OK rather than outright calling this out as "redpill bullshit" can you refute it with any substance?

Chaos is indeed deterministic but only in the first few iterations. The limitation we have is computational power. The further into the future we go, the hard it is to determine the outcome.

The best example of this would be weather forecasting. A lot of weather bureaus can perfectly forecast the next 24 hours but become increasingly difficult to get it right the more into the future you get them to forecast.

Welcome to Feudalism 2.0 - your new overlords are no longer the Duke of Aquitaine, Lorraine or Normandy but Google, Amazon, Facebook and Apple.