The EU in its current form is mostly about markets. It routinely pushes for the sacrifice of government monopolies to the altar of the free market (see for a recent example the french train network). Most of its regulations are to ensure a level field for a balanced market.
Hell it pushes for free markets even when it makes very little sense (the eu electricity market and its weird idiosyncrasies are an artifact of that)
It basically bans member governments from printing money and imposes very strict limits of 3% GDP on government deficits. For reference the US deficit was 5.9% gdp this year, Almost twice as much. this greatly limits government control over the economy.
> imposes very strict limits of 3% GDP on government deficits.
You might want to check that information. This very strict limit is only enforced on selective EU countries like Greece for example.
France has had routinely yearly deficits above 3% in the last 10 years and has never been worried one bit about it.
For the record the French deficit was around 5.4% this year and it is set to increase again next year as the parliament is completely blocked and a budget compromise cannot be reached.
Even the so called debt ceiling defined in the pact of stability is mostly ignored. Italy and France are both well above the 100% debt to GDP ratio when the treaty says that every country within the EU should be at or below 60%.
> It basically bans member governments from printing money
It only bans the ones that have adopted the Euro. The countries that have declined to adopt the euro are free to do as they please more or less.
The euro countries though may not be able to print money, but they just get the ECB to do it for them via quantitative easing which has been used since 2008 and only recently stopped when the interests rates started climbing after the pandemic.
As someone who worked on transformer-based diffusion models before (not for language though), i can say one thing: they're hard.
Denoising diffusion models benefited a lot from the u-net, which is a pretty simple network (compared to a transformer) and very well-adapted to the denoising task. Plus diffusion on images is great to research because it's very easy to visualize, and therefore to wrap your head around
Doing diffusion on text is a great idea, but my intuition is it will prove more challenging, and probably take a while before we get something working
In between the beds?? Does that mean the shower was right in the middle of the room ? So that it would be impossible to place a double bed ? This is the weirdest part to me
I tried that concept whith my ergodox when i had an arm in a splint, but i couldnt quite get my brain to wrap around it. I could type on the right key, but not press the mirror/switch key at the right moment.
What would have made it easier is if it could infer the right key like an autocorrect
Python has static typechecking which, while not perfect, works pretty well as long as you're not actually trying to circumvent it (or manipulating things "too dynamically")
Python does allow you to put anything in annotations. ( pep 3107 that defines type annotations says that explicitly [1]).
But it also defines a type checking annotations, which is a specific convention for using pep 3107 annotations. Type annotations were introduced in PEP 484 and updated in a lot of subsequent peps. The python typing system is fully specified in [3].
It does have several implementations, although the reference implementation is mypy
This feels like something that should be opt-in, not opt-out. It feels trivial to have all clients that support it send a header stating they do, and it is ridiculous that the default is to allow sending reacts to clients that don't support them
It's a embrace-extend-extiguish play like the old days. Add a 'feature' that doesn't technically break the rules (or only does a little), get your users used to it (by making it the default, opt-out, etc) and hope that your users will pressure people not using your product to move. "What do you mean you didn't see my email reaction? That's the best feature in the whole world. You should really switch to outlook, etc.". See: every M$-only feature in IE.
How is this argument not just “no one should ever implement new features”?
I don’t really care for the Outlook reactions and find them out of place, but this implementation doesn’t break anyone else. It’s also exactly how Apple implemented reactions being sent to SMS recipients.
Yes, we got the "you're just a luddite that hates progress" sophistry from you guys in the IE days (and before). "We're doing the same thing as Apple" isn't a particularly persuasive counter. I always appreciated Balmer in that that he didn't waste time bullshitting anyone that he was trying to create walled gardens for M$ products by cooping standards.
This feels like something that should not exist, period. For any email important enough to actually send, asking people to guess what a single-emoji "reaction" actually means is a recipe for bad communication.
"And hence invertible" <- does every output embedding combination have an associated input ? Are they able to construct it or is this just an existence result ?
I don't think they're claiming surjectivity here. They're just saying the mapping is injective, so for a given output there should be a unique input to construct it.
Usually we still ask for statistics to be at least valid (i.e. have a significant signal under a null hypothesis). This paper doesn't even do that. It's like claiming no humans have been to the moon and then "verifying" this by randomly asking a million random strangers on the street if they've been there.
