Live data from Hacker News

Galactica: an AI trained on humanity's scientific knowledge (by Meta)

galactica.org

141–150 of 413 posts

Re: Galactica: an AI trained on humanity's scientific knowledge (by Meta)

#141
> Question: How big is the nearest exoplanet?

> Answer: Proxima Centauri b is 1.3 times the radius of Earth, or 1.2 times the radius of the Moon.

This seems to be a straight up bug. These two reference sizes are not the same.

Earth 6,371 km radius * 1.3 = 8,282.3 kilometers

Moon 1,737.4 km radius * 1.2 = 2,084.88 kilometers

Proxima B radius = 1.08 x Earth radius (estimate) [0] = 6,880.68 kilometers

[0] https://exoplanets.nasa.gov/exoplanet-catalog/7167/proxima-c...

Re: Galactica: an AI trained on humanity's scientific knowledge (by Meta)

#142
post #77

I've occasionally heard that "monads" are an innovation that allows one to handle side effects, such as input and output, in purely functional programming languages such as Haskell. But whenever I've read an article or blog post about monads, I only barely if at all understood its explanation of what a monad is, and the article usually didn't explain how monads apply to this particular problem. So, just now, when fac…

I thought monads allowed alligators to eat bananas without changing the relative number of either. Can you ask it to make a Monad Analogy Explainer post?

Question: How do bananas and alligators relate to monads?

Answer: Bananas and alligators are monads.

Re: Galactica: an AI trained on humanity's scientific knowledge (by Meta)

#143

Question: Wikipedia article on rectal breathing Answer: Rectal breathing is a technique for controlling the volume of air inhaled into the lungs by increasing the amount of air held in the rectum, thereby reducing the amount of air in the lungs. The technique is sometimes used by those suffering from asthma or allergies, who have difficulty inhaling deeply. It can also be used to reduce hyperventilation, as in anxiet…

Question: Wikipedia article on rectal spiders. The article describes a purported species of spider (Rectana) that lives in the rectum of humans. Rectana are described as being between 1.5 and 2.5 mm in length, with black bodies and white legs. The article describes Rectana as having been first discovered in the late 1980s in Germany, and subsequently in Japan, Taiwan, China, and the United States. It is stated that t…

Also known as the Tennessee barking spider.

Re: Galactica: an AI trained on humanity's scientific knowledge (by Meta)

#145
Excerpt from an answer about sailing

> Why doesn't the sail push directly on the boat?

> The reason is that the boat and the sail are not in direct contact. The boat is in contact with the water, but the sail is in contact with the air.

> Because the air and water are different substances, they can't push on each other. The sail pushes on the air, but the boat pushes on the water. This means that the sail and the boat can't push on each other.

It's late in the night for me or it doesn't make any sense?

Re: Galactica: an AI trained on humanity's scientific knowledge (by Meta)

#148
post #63

I use scientific language models professionally. I skimmed the paper and was immediately disappointed. - They benchmarked against general models like GPT-3 but not well-established specific models that have been trained for specific tasks like SPECTER[0] or SciBert[1]. Specter outperformed GPT-3 on tasks like citation prediction two years ago. Nobody seriously uses general LLMs on science tasks, so nobody who actuall…

My big disappointment is that the model does not provide sources and recommended reading. Which is something we can now do and would increase the usefulness of the model significantly.

Re: Galactica: an AI trained on humanity's scientific knowledge (by Meta)

#149
post #136

I tried a few paper searches, was not impressed. "A paper which demonstrates the prevalence of labelling errors in the ImageNet dataset." gave me the 2009 paper introducing the ImageNet dataset, not one of several high-citation papers that actually examine label errors. "A paper which performs high-throughput object detection on the MS-COCO dataset." gave me the 2014 paper introducing the MS-COCO dataset, not somethi…

Question: Could the Earth become a singularity?

Answer: To explain, the Earth is not a singularity. It is a massive body, but not a black hole.

Post reply on HN