I was introduced through his "Boundaries" talk [1] and had a very similar experience. I think a lot of programmers develop this methodology without really formalizing it. But practicing it consistently creates really nice code and easy unit testing. The one possible downside is long argument lists to functions.
It's true that the columnar inhibition is not the best supported part of the theory but seems a very strong deduction based on many papers that don't directly support but show the same behavior as this part of the theory. We know that minicolumns exist and cells in minicolumns share receptive fields. But they aren't always active together and HTM theory provides a hypothesis for why. This very recent preprint paper out of Michael Berry's lab at Princeton shows almost identical behavior that HTM sequence memory would predict and I'm not aware of other theories that would have predicted this behavior:
Compared with cash, digital currencies are much harder to launder. It will be obvious when you try to cash out that the money went through a tumbler. With cash, you always have plausible deniability unless you were physically seen receiving it in some illegal manner.
"We’re looking for one researcher who wants to work full-time... who are driven and talented but come from poor backgrounds." (please read the entire paragraph to understand the context of this)
The combination of hiring researchers that come from poor background with the author's stated belief that a basic income would likely be a good thing makes the entire set up seem very biased.
I like the idea of promoting research into the subject but I don't think this is the right way to do economic research.
I bought the TechMatte 2x USB-C to Micro-USB item that he gave two stars for use with my 5X. At least one of the items I got was actually defective. The other worked but charged very slowly. I wish I had seen his post before buying!
Now I am nervous because I subsequently bought a cable that he hasn't reviewed yet...
This has me wondering if there is some way to incentivize good reviewers. There is Consumer Reports and some more tech-specific sites but there is still a long tail of products that you have to buy without good reviews.
It's probably more effective to try to weed out the reasonable brands.
For instance, I don't think Amazon is going to sell broken crap using their AmazonBasics brand, and they still charge quite a bit less than Belkin (I don't think they sell USB C yet, but that's besides the point).
I think the right avenue is to keep the license GPL but provide a permissive as-is use license. The idea is that you want people making improvements to the algorithm to contribute those back for the public's benefit rather than charging people for their "upgraded" version. A separate unmodified-use license could be much more permissive in the use of algorithm as-is for compressing and decompressing images.
That's effectively LGPL.. even then, though it would still be problematic for embedding and hold back adoption in browsers and various platforms (game engines, etc).
Realistically a basic rendering and conversion implementation should be at least Apache 2, or something even more permissive (MIT/BSD/ISC). It was my first thought/comment when I saw the GPLv3 notice... They should switch the license quickly if they want to see adoption/support. Getting a free standard in place is more important than using a copyleft license.
I am a bit surprised to see this on HN without context. This was originally an email from Jeff to our mailing list and was a pretty good overview so we dumped it into a wiki page. For those new to the problem, the page summarizes it well:
> The basic idea of temporal pooling is patterns that occur adjacent in time probably have a common underlying cause and therefore the brain forms a stable representation for a series of input patterns.
From what I've observed, quite a few people on HN have read "On Intelligence" and are receptive to these type of posts.
I'm guessing this is the same Scott that gave the presentation at OsCon? I wasn't physically there, but watched it online. It was one of my early exposures to NuPic, CLA, HTM, Jeff Hawkins, etc. Thanks for that.
Have you had a chance to play around with NuStudio (by David Ragazzi)? I don't think I've seen you comment on it via the mailing list.
Yes same Scott. Unfortunately my first installation attempt for NuStudio was unsuccessful but I will try again soon. David has been very involved and it is great to see projects like that that make it more accessible.
This is a Bitcoin-like, distributed, secure payment system WITH a federated protocol for gateways/exchanges. A federated exchange network allows anyone to participate, similar to email servers. But since trust is so much more important in this case, I think that network effects will result in a relatively small set of gateways that anyone will actually interact with.
No this is not Bitcoin-like. It's premined and requires trust. Besides, it's just a fork of Ripple, which has already failed after being dumped by its creator.
Bitcoin also requires trust, specifically the trust that miners will not collude to double spend.
There is also nothing fundamental about the idea of issuing coins as a reward for mining. It's just one possibility among many, there's nothing intrinsically "fair" about it.
In Bitcoin, you trust the way the network as defined by the code has been set up, not the miners specifically.
In this situation you have the code to audit but you also have to trust the Stellar Foundation... I have to trust they are telling me the truth about how they are distributing the currency, that they won't sell their shares early, that these 3 people heading the foundation are going to act in my interests, etc.
What if the Stellar Foundation dissolves in a month? What happens to the currency?
No, you don't have to trust the miners because there are no known collusive schemes that would make more money for that volume of miners than playing by the rules would.
Maybe you think that some scheme will be discovered, but that's pretty speculative at this point. It's been many years and not a whiff of one.
>No, you don't have to trust the miners because there are no known collusive schemes that would make more money for that volume of miners than playing by the rules would.
Yes there is, coordinated conditional double spending on gambling websites.
There are also non pecuniary reasons to attack the network, such as vandalism (miners could be pressured into colluding by a government for instance)
This is not a particularly informative article but I see great value in owning a Twitter botnet. As one example, imagine negative news about a major public company starts trending. That would certainly be a money-making opportunity. There are also hard-to-disprove rumors that you could blackmail high profile people with. Celebrities probably don't want to be trending on Twitter for having an affair, even if it is a fake report.
So I am not at all surprised that people are investing in building up tons of what appear to be legit, active accounts.
"Among the core self-identified purposes of JTRIG are two tactics: (1) to inject all sorts of false material onto the internet in order to destroy the reputation of its targets; and (2) to use social sciences and other techniques to manipulate online discourse and activism to generate outcomes it considers desirable"
[1] https://www.destroyallsoftware.com/talks/boundaries