subroutine

What happens to the water used by data centers? Is it still clean, just warmed up?

I see where you are coming from. I imagine LLMs will follow the same feedback lifecycle as other groundbreaking tech. Most legitimate iPhone 1 complaints came paired with a few positive messages to the devs and engineers. Years later, after countless improvements to the hardware and UI based on user feedback, now it's just complaints.

Based on the amount of low-effort criticisms out there, clearly nobody is afraid of "big AI". Tao seems careful about his critiques of AI because low-effort off-the-cuff criticism can lead to all sorts of future problems/hypocracy.

> you need to expect appropriate legals consequences for this sort of negligence

I might have missed it, but did the agents do something illegal? Or do you think that what the agents did should be considered illegal?

But Claude Opus 4.6 is not really practical. Taalas' process seems targeted for edge models. Their proof of concept model, for example, is a heavily quantized version of Llama 3.1 8B and even then they acknowledge their custom 3-bit/6-bit representation causes model quality degradation.

Taalas is going to have a tough time putting a trillion-parameter model on one conventional die. Their HC1 die is already near the maximum size that conventional lithography can expose. They claim they could partition the model across many chips, but I'm not sure if they have tested this process or what it means for compute. The basic storage arithmetic is unforgiving: for a one trillion parameters model at four bits it will take 50–100 chips. To service a sizable customer base will take thousands of 100-chip fabs.

That all said, I'm bullish on this technology, and look forward to seeing it evolve.

You prefer to browse amazon using a mobile web browser over the native Amazon app? Why?

Why would this be more likely? The current remotes already have a Netflix button, a Hulu button, a Sling button and a Disney+ button. Roku sells dedicated buttons on their remotes to the highest bidder.

Oh shit, you got me! In all seriousness, you are not doing yourself any favors if this is how you typically debate a topic. I expect more from a HN netizen.

Until you look; then it collapses to one or the other.

In theory, quantum computers could provide dramatic speedups for certain linear algebra operations (eg. matrix inversion, eigenvalue estimation). The catch is that many NN training algorithms need all the data to be stored in qRAM so the QC can access matrices efficiently. Loading in a massive dataset will likely be more difficult than the computations, eliminating the quantum advantage. This is analogous to having an extraordinarily fast processor attached to a slow af hard drive inside a neutrino storm.

A good thing for anyone who isn't a Nazi simp or a fan of authoritarian takeovers and imperial rule.

The U.S. Navy did not blockade Japan or “prevent oil being shipped” by closing the seas. The U.S. imposed an oil embargo and froze Japanese assets after Japan expanded its war in China and moved to invade other pacific countries. Surely you can understand why that was a good thing.

Why would you post such nonsense given how easy it is these days to determine bullshit? By the time of Pearl Harbor, Japan was formally aligned with Nazi Germany. Japan, Germany, and Italy signed the Tripartite Pact in Sept 1940 creating the Axis alliance. Pearl Harbor happened in Dec 1941, so Japan had been formally tied to Germany for more than a year.

“The American navy closed international waters.” Not in the Pearl Harbor context. Before Pearl Harbor the U.S. was not conducting a naval blockade of Japan that closed international waters. The U.S. cut off Japan from US oil in July 1941. That is not the same thing as the U.S. Navy closing the Pacific.

“The USA blockade was the reason Japan attacked Pearl Harbor.” False. Japan attacked Pearl Harbor because it wanted to neutralize the U.S. Pacific Fleet while Japan seized the "Southern Resource Area”, especially oil-rich East Indies, Malaya and other regions in the pacific. The U.S. oil embargo might have played a small factor, but that wasn't a US-only thing; various countries were increasingly unwilling to sell oil and other resources to Nazi-aligned Japan while they were attempting to conquer China and most of the Southeast Pacific.

And let's also not forget that keeping wildlife, an amphibious rodent, within the city isn't legal either.

Why would disinformation about a technology used to recover a pilot have any bearing on public socio-political views about the war? Eternal September is becoming increasing real on HN.

Because not everyone has a transmitter, but everyone has a heartbeat?

Though in this case the pilot likely had a transmitter and that's exactly how they found him.

Because the NY Post ran an article that said

"The CIA used a futuristic new tool called “Ghost Murmur” to find and rescue the second American airman who was shot down in southern Iran, The Post has learned. The secret technology uses long-range quantum magnetometry to find the electromagnetic signal of a human heartbeat"

Note, I agree that it was probably some novel beacon technology. Just answering your question about why people are debating whether it was a device that could detect a human heartbeat from long range.

Are you asking if the 10 seconds it takes AI to generate an image is more costly to the environment than a commissioned graphics artist using a laptop for 5-6 hours, or a painter who uses physical media sourced from all over the world?

Has anyone tested the system from the other end... sending a prompt and getting a response?

Copy?

Gmail is free. How much customer support resources should someone reasonably expect a company to dedicate towards their free-of-charge services?

Some items on thingsverse provide .obj files; like the king in this chess set...

https://www.thingiverse.com/thing:1078513/files

or this army tank...

https://www.thingiverse.com/thing:4618182/files

(n.b. under the main image viewer click the "files" tab to explore individual files/extensions)

Oh duh, you are absolutely right. I thought you were saying they never released GPT-2.

I'm confused. From what I can tell GPT-2 was released in 2019...

https://huggingface.co/openai-community/gpt2

https://github.com/openai/gpt-2/blob/master/model_card.md

From the Wikipedia article...

For planning Operation Epic Fury, the US military utilized the Maven Smart System, an artificial intelligence software designed to streamline the targeting process and greatly reduce the amount of personnel involved in it. Capable of producing 1,000 target packages in one hour, with the use of the system the US military said it had struck 6,000 targets in Iran during the first two weeks of the war.

...it goes on to say...

The [NYT] inquiry suggested that the school was likely targeted due to outdated coordinates provided by the Defense Intelligence Agency

Advanced rockets bolted onto mainframes guided by data from Palantir.

https://en.wikipedia.org/wiki/Project_Maven#Technology

They finally did release 2.0 under the MIT license. That was the last version (a 1.5-billion-parameter model) they would release open source. GPT3 for comparison has 175 billion parameters.

At 20 min per task you might as well code it yourself. Bill James needs to write a book on saber-metrics for LLM benchmarks.

See French Decimal Time:

https://en.wikipedia.org/wiki/Decimal_time

Not to be confused with Metric Time:

https://en.wikipedia.org/wiki/Metric_time

Timekeeping units of measurement:

https://en.wikipedia.org/wiki/Unit_of_time

It's such an odd request to make something less enjoyable. If the EU wants a time limit on app use they should just impose it themselves.