

I’m saying there is no “big leap” necessary. As the paper that introduced the transformer said, attention is all you need.


I’m saying there is no “big leap” necessary. As the paper that introduced the transformer said, attention is all you need.


Transformer is useful for damn near anything. At the end of the day, what we consider intelligence is the ability to predict what comes next, whether that is what our senses will tell us next or what the next hypothesis to test should be based on the data we have seen so far.


Solving famous unsolved problem: https://www.scientificamerican.com/article/ai-just-solved-an-80-year-old-erdos-problem-and-mathematicians-are-amazed/
Drug discovery (AI can’t do anything about the time needed for clinical trials): https://www.nature.com/articles/s41591-024-03434-4
Assisting the blind: https://www.acb.org/i-am-blind-self-driving-cars-waymo-give-people-me-independence


Uh, no it’s not.
It is. As a result of the Epic Games v. Google, Android builds with the Play Store are required to allow users to install apps without any warning at all. They obviously can’t allow any app to be installed without a warning because this would be a boon to malware authors, so this is now enabled with verification. You can now even share apps you build with your friends without requiring them to go through an unverified apps flow with a scary warning. Additionally, Google is not allowed to take a revenue cut from those installs.
You’re confused because the install process for apps that are not verified (a path that didn’t exist before at all) or installed from a system app store has changed. This now has to be done with adb, which takes effect immediately, or via an on-phone process that takes a day to complete. Once it is done, this setting is copied to new phones, so the process actually becomes easier for most people who do this because they don’t have to go through the process repeatedly.


This change is the opposite. It makes it possible for a user to install the Epic Games Store from their website without seeing a scary warning, and Google won’t get a cut of any of the revenues from that store. The same with any other company. Netflix can now offer their app from their website, and people can install it without any warning, and Netflix won’t have to send any revenue to Google for people who subscribe in the app.
Removed by mod
And the “history” of the event you were taught was a fantasy cooked up by McCarthyist acolytes to spread atrocity propaganda.
The history I was taught is backed by documentation. The history you were taught ignores the documentation and takes the word of a government that tried to destroy any documents that existed of the event.
Took me all of 1 second to realize that the massacre happened after the 1940s.
Then it took me all of an additional minute to read what happened after the 1940s in the Wikipedia article:
“The Chinese Communist Party re-organized the Red Cross organization in Beijing and was admitted to the International Federation in 1950.”
The “history” you will be taught in that course won’t include the reports of the journalists who were there at the time and whose film was confiscated. The famous tank man picture only survived because the journalist who took that picture hid it in a toilet.
The Chinese Red Cross estimated 2,600 fatalities. That’s a lot of fishermen.
ROC doesn’t claim to legitimately rule China anymore. That was just KMT’s delusion, and they’re no longer in power.
I think a reasonable game could be played with those rules, given how quickly goals are scored and how hard it is supposed to be to catch the snitch. It’s just that it didn’t make sense at all that Krum was celebrated. Catching the snitch was worse than scoring an own goal in soccer because it directly and immediately caused his team to lose the match. The rioting of the death eaters after the match is understandable, but the way everyone else behaved towards this obvious fraud is not. The Ministry should have started a match fixing investigation.
You might enjoy Vinge’s Zones of Thought series.
In The Goblet of Fire, Ireland beat Bulgaria despite Bulgaria getting the snitch. The problem with the snitch isn’t that the team that gets it automatically wins but that this particular match didn’t make sense because Bulgaria knew that getting the snitch would cause them to lose, so they would have instead focused on preventing Ireland from getting the snitch while they tried to get within 150 points.
The American Immigration Council is a better source than Al Jazeera for describing what ICE might be doing that is wrong, so I don’t consider that a valid criticism of this summary.
This is common practice. If you look at Google Maps in any other country, the name will not have changed. In the U.S., it shows the legal name within the U.S., which was updated by H.R. 276.
I asked about the Strait of Hormuz, and it summarized information from Al Jazeera, the BBC, and Wikipedia. It added the following disclaimer:
Disclaimer: This information is based on reports from April 18–20, 2026, and the situation is highly volatile.
This seems reasonable.
When I asked about ICE illegal operations, it summarized and linked to the American Immigration Council.
This is not as good. If I’m asking about something done right now being illegal, I would like to see ongoing cases challenging the legality of the actions. I’d hesitate to call that bias instead of just bad results though.
I also didn’t get Internet connected thermostats until the utility company added demand response discounts. It’s really a smart grid technology. This does mean that it should be secured as such, otherwise it’s another vector to attack the power grid (set all thermostats to maximum and cause blackouts). Regulations haven’t caught up.
In much the same way as human thinking is the second best (and soon third best) solution to any problem. The point is that an LLM can come up with the best solution and use it.
Obviously not — they’re not going to make claims beyond the results they achieved in the paper. It was, however, obvious to everyone who read the paper that all of what we consider thinking could be derived by clever application of a sequence model, and all those papers that came after were results achieved by teams doing the obvious thing.