The demo is showing System One producing its output in parallel very quickly and for little cost compared to an LLM generating its answers token-by-token. The "noul" type is used to evaluate a yes/no question and return the probability that the answer is yes.
So this demo is showing System One offering much more nuanced responses and specific probabilities compared to an LLM's more crude responses (e.g. LLM shows "true" or "false" compared to "0.9" or "0.07" probabilities that the answer to some question is true).
Hey, that's awesome and nevermind me. I just got stumbled by their UI.
There's probably a fair argument about how discoverable these are (especially given their labeling as "All Issues" and "All Pull Requests") but that tip is quite helpful to me personally. Thanks for sharing it, I really appreciate it!
There may be some magic they do to better optimize within-user-searching. It's something that they could hide in implementation details so we can't be sure unless they spill the beans but it's feasible - especially with the default search parameters they're using.
I'd still love something a bit more obvious and intuitive but if it's just a UX failure that makes me feel a lot better.
> There may be some magic they do to better optimize within-user-searching. It's something that they could hide in implementation details so we can't be sure unless they spill the beans but it's feasible - especially with the default search parameters they're using.
I would have believed this until last week when they had a little banner informing me that I might not see all the PRs but that they really were there and that I could use such-and-such API to find them myself.
If the direct API existed and was working, but the web UI wasn’t seeing the PRs, then presumably it meant that the web UI was not optimizing the default trivial search query to use the direct API.
> if a website cannot reliably detect a VPN user's true location and the law requires it to do so for all users in a particular state, then the legal risk could push the site to either ban all known VPN IPs, or to mandate age verification for every visitor globally.
Clearly anyone slightly sophisticated can bypass restrictions like this. A quick search reveals https://github.com/shadowsocks. This only harms regular users who might benefit from privacy. The dystopia levels continue to rise...
Ah your comment finally makes me understand the premise of MMT[1], which seems to presuppose that the US will always have this special status. Makes the current administration's geopolitical recklessness even more terrifying.
Taxes, if not quite the price of civilisation, do give citizens a reason to care about efficient and effective government. Severing that connection, and leaving large chunks of the electorate as mere recipients of state largesse, risks deepening America’s political dysfunction.
We can't have nice things without paying for them. People who believe they are self-sufficient seem to ignore all the public infrastructure that keeps society and the economy moving (e.g. roads, emergency response/firefighters, schools, parks, libraries, etc).
Imagine how much more entrepreneurial people could be if taking big big financial risks didn't have dire consequences like not having access to health care.
No one loves paying taxes, especially when you don't agree with ways it's spent, but that means we need to fix politics and spend money better rather than denying that society needs financial contributions from almost everybody to function.
But what they do use a lot of is a reliance on educated employees, a wealthy customer base, a peaceful business environment, courts that enforce contracts, a stable currency, regulated and policed financial system, et cetera.
All of the above is highly under-rated and acknowledged IMO.
Roads and schools have been a net negative for American society. Massive amounts of public money are spent on making problems worse for no gain while simultaneously bloating spending further. (e.g. Smokey Bear, road build-out, bailouts).
Citizens shouldn't have to care about the efficiency of the Government, it should be allowed to fail like any other company. The economy wants off Uncle Sam's lap.
> if run at full strength for a year, Colossus [xAI's new datacenter] would use as much electricity as 200,000 American homes. When fully operational [...] this facility and two other xAI data centers nearby will require nearly two gigawatts of power. Annually, those facilities could consume roughly twice as much electricity as the city of Seattle.
> Even conservative analyses forecast that the tech industry will drop the equivalent of roughly 40 Seattles onto America’s grid within a decade; aggressive scenarios predict more than 60 in half that time.
> [...] by 2030, U.S. data centers will consume more electricity than all of the [USA]’s heavy industries [...] put together.
Basically it boils down to the US meeting this electricity need using natural gas and fossil fuels (with nuclear as an eventual goal) and China moving at a similar pace using renewables. Of course the AI companies on the west coast don't have to experience the pollution such energy demand brings to the communities nearby the power and data center facilities.
The implication from the LinkedIn post (to my eyes at least) is that they believe they have a case for canceling X Corp's claim to the Twitter trademarks. And then the landing page at twitter.new certainly implies that they plan to start their own social network/app called Twitter.
The demo is showing System One producing its output in parallel very quickly and for little cost compared to an LLM generating its answers token-by-token. The "noul" type is used to evaluate a yes/no question and return the probability that the answer is yes.
So this demo is showing System One offering much more nuanced responses and specific probabilities compared to an LLM's more crude responses (e.g. LLM shows "true" or "false" compared to "0.9" or "0.07" probabilities that the answer to some question is true).
reply