Interesting but it feels like giving an agent the HTTP protocol documentation and a raw socket tool instead of curl, to make requests. Very wasteful in tokens and ultimately computation.
This is of course because we are limited to using agents interactively to avoid paying the API rates.
An agent harness with this integrated would be much more efficient.
I have been made aware of ReFS thanks to the "Dev drive"[1] technology in Win11. ReFS is apparently particularly suited to store source code and code artifacts for dev workloads.
I have been using it (dedicated DevDrive partition) for some time but performance improvement are subtle and not easily visible without measurements. Which I haven't done.
I agree about the importance of better infrastructure but I had to check this 50% loss number.
Water distribution efficiency in France was actually 81.5% in 2021.[1]
Which means 18.5% is lost to leaks.
The figure often given is 20%, maybe taking into account the error margins of the studies.
According to another study from 2019 [2], it's actually similarly or more efficient than other rich countries in Europe (losses of 21% in the UK, Belgium 27%, Italy 34%).
The author of the study says that according to experts reaching efficiency above 85% is extremely difficult. [2]
Mandating that the takedown go through this agency means states won't be able to hire private companies specialising in takedowns (and faster). Takedowns take way longer, which hurts the children.
This is a very common complaint on the internet, the introductory paragraphs of recipe blogs.
Yet, when I am looking for a recipe, I always look for the blogs.
Why ? Because the alternative is a commercial website full of ads and trackers, trying to upsell me on paid features on every page.
Yes I have to scroll past an uninteresting paragraph, but it's much better than the alternative.
This new website might be more respectful of my privacy now, but it is yet to be seen what will change when/if they need to monetize it.
Nice game, but the server may be overloaded right now.
It is possible to be in 2 games at once (and losing both).
A bug made my game end as if the opponent had quit, I went back to the menu and started another one. Made one or 2 moves and then I was back in the game I had just quit, where my opponent was still playing.
Of course I was AFK for 1 turn (playing the other game) so my position was revealed. Then I was switched back to the other game where I had also appeared AFK for a turn.
Not impressed, it can deduct gender, age and accent (with biases) but calling the portrait 'accurate' is clickbait.
All the use-cases we are told about (like police sketch) are totally irrealstic. The portraits look "realistic" but they do not look like the model . They look like a different person with the same ethnicity, similar age and gender. With biases, again...
Given Google is not in financial difficulty and these employees were not at fault of anything, this layoff would have been illegal in France.
If these employees were not contractors or on a short-term contract, this layoff could easily be contested and Google would have to pay up to 1 month's pay per year worked at Google (minimum 3 month's pays)
Google could have negotiated with the employees to end the contract at a better price.
Yes ! Computer Emergency Response Teams (CERT)[1] exist in most countries and publish security advisories as newsletters or RSS.
e.g. CERT-EU security advisories [2]
But there are so many softwares and exploits that the signal to noise ratio is low if you are not in charge of a big IT infra.
When you read this with the classical crypto-skeptic arguments in mind, it's hard not to "cringe".
How can "Rare Pepes" and "Crypto-kitties" be the reason you invest in a teghnology ?
And yet those investments must have been highly profitable.
I am still a crypto-skeptic but I wonder if in the future my tendency to over-analyze (based on intuition and generalist tech knowledge) will make me overlook a truly revolutionary innovation.
Am I wrong about crypto ? Will the metaverse be this tech I overlooked ?
Maybe this article is right and I should keep it simple. Trust the hype and stop worrying ?
I visited improving.duckduckgo.com and they are linking to this article[0] that explains how this domain hosts an "anonymous tracking pixel" (that respects DDG privacy policy) to collect data on how DDG is used (feature usage, A/B testing etc)
But it does not tell how it could end up on other domains. Are you sure it is not part of the DDG result page or extension ?
By the way, you are including a lot of useless files and binary files in your git repository.
(.egg folder, pycache, pyc files). They make your repo very heavy.
Take a look at this example .gitignore : https://github.com/github/gitignore/blob/main/Python.gitigno...
This is of course because we are limited to using agents interactively to avoid paying the API rates. An agent harness with this integrated would be much more efficient.