A new law in Scotland threatens you with seven years in JAIL if you misgender someone.
GLAD @jk_rowling calls this a threat to free speech.
Here is how she protested:
RFK Jr. is solid when it comes to Israel and the Palestinians.
Definitely worth watching the full 6 minute clip.
Highlights:
-“Listen, if Mexico attacked us and we built a fence. Would you blame us for caging in Mexico? Everything in your mind is telling you to blame Israel, instead of blaming Hamas.”
- “First of all, Israel has no obligation. Israel built 3,000 houses and gave them for greenhouses. That would have made Gaza completely food self-sufficient, gave it to them as a gift. Offered to rebuild the port of Gaza to make it the Singapore of the West. Hamas said, ‘No, we don’t want you’re money. We don’t want you’re ideas.'”
-“And what did they do? The international aid agencies have given Hamas, have given Gaza more than 10 times per capita what we gave to rebuild all of Europe after the Marshall Plan,” he continued. “They’ve gotten $8,300 per capita, every person in Gaza. We rebuilt Europe with $621 per capita in Europe, and we rebuilt it. What did they do with that money? Instead of using it to make this, you know, Gaza this beautiful country with white sand beaches. It should be a paradise. Hamas said, ‘we don’t want that.’ They take virtually all the money and they steal it. So the top five, the top five leaders of Hamas are billionaires. Ismail Haniyeh has $5 billion, according to Forbes.”
-“It’s not Israel’s fault that Gaza is poverty stricken. Gaza should be one of the wealthiest states on the Mediterranean.”
-“Krystal. Krystal. Krystal, why are you insisting on blaming Israel? Why don’t you blame Hamas? You know what most of our tax dollars have gone for? The Iron Dome, which is a way of not invading Gaza. Our country and Israel, for 16 years, have expended this huge amount of money to try not to go into Gaza, while Gaza sent 2,000 rockets a year, suicide bombers, and of course Israel —”
@fchollet so we have to add a-priori bias towards a specific "something" we want it to learn. And when we do, we influence the "conversion ratio". And we can make that ratio arbitrarily big, by having an agent that knows apriori almost all of that "something" we want it to learn,
@fchollet Another way to say it - "how good it learns" depends on "what we want it learn".
Without at least some a-priori bias towards "what we want it to learn", learning is impossible, because the hypothesis space is infinite and we cant narrow it down no matter how much data it sees.