jeffreyames

A system that’s more robust to meddling than you’re imagining

You’re imagining a 10x payout for sabotaging a trial. That would be a trial with about 90% odds of succeeding when most trials are likely to fail. And this isn’t a high school science experiment a single janitor can easily sabotage. Any drug with high odds for approval has already shown significant evidence of efficacy and safety. Tampering would not shut down the whole development process with hundreds of millions poured into it.

It can easily detect repeated failed attempts at a non obvious problem such as incorrect documentation or missing run flags

Precision and ahead of schedule doesn’t accurately describe these projects once they reached a certain level of cost and complexity over a half century ago. I recommend reading Prophets of War Lockheed Martin and the Making of the Military-Industrial Complex to get a sense of how remarkably efficient spacex has been in comparison.

I imagine for a quadriplegic moving around a 3D world in first person by thinking must feel great. I wonder if the thought that triggers the movement feels more like thinking about walking or thinking about pressing a button on a controller.

Gpt4’s response to the question using the exact wording

“To find out who wins, we can calculate the time it takes for each car to complete the race.

For Car A: - Speed = 10 mph - Distance = 100 miles - Time = Distance / Speed = 100 miles / 10 mph = 10 hours

For Car B: - Speed = 5 mph - Distance = 50 miles (because of the 50 mile headstart, it only needs to cover 50 miles) - Time = Distance / Speed = 50 miles / 5 mph = 10 hours

Both cars would finish the race at the same time, so it's a tie.”

Many people have internet speeds faster than the ps5 blu-ray drive read speeds.

You can try Raytheon https://en.wikipedia.org/wiki/Active_Denial_System

They said

> we've heard all your feedback about GPT4 getting lazier! we haven't updated the model since Nov 11th, and this certainly isn't intentional. model behavior can be unpredictable, and we're looking into fixing it

This contradicts itself each sentence and sounds as if it was written by chatgpt itself in lazy mode.

They continue to set an incredibly high bar from story to engineering with every major release. It’s full of talent pushing the state of the art. Dead, utterly disgraced doesn’t describe the company. Stories about employees feeling exploited is more of a successful, topical company thing than a Rockstar thing.

High performance on a single task requires simultaneous computation and communication between nodes. If there's high latency between nodes, such as between nodes in different data centers, the communication costs can't be masked by computation.

10k H100 chips is considered a very large cluster. The third fastest supercomputer in the world is Microsoft’s eagle with 14k H100s https://www.top500.org/lists/top500/2023/11/

This could be used as a consciousness override, replacing your own thoughts in realtime. I can think of some good applications and many terrible ones

Not just secrets but deliberate misinformation in the name of national security

There are many unselfish people who don’t like to throw good money after bad, who think that the current solutions are so ineffective that even if the entire wealth of every billionaire went towards homelessness the needle would barely move.

> I'd probably have no qualms about stabbing someone who I knew was trying to introduce my child or SO to hard drugs.

Have qualms. You stabbing someone will be worse for your child than hard drugs.

Totally agree though a lot of video conference software mysteriously doesn’t allow you to disable this.

A company I was interviewing had a document they shared with applicants with interviewing guidance. One tip was to look directly into the camera when speaking so the interviewer feels you are present and attentive. Imagine two people looking at a camera instead of each other in an absurd attempt to make video chat more natural.

Check out monitorcontrol. Think it works over usb-c in your case but not positive. https://github.com/MonitorControl/MonitorControl

I’m fluent in English and I find autocomplete helpful. More advanced tools like Writefull that can help paraphrase or structure larger blocks of text have a lot of potential. These next generation autocomplete tools already are and will be increasingly useful to experts.

The tools can reduce the time and mental energy spent writing boilerplate and provide consistent structure to code and comments. This allows me to operate at a higher level of abstraction instead of worrying about details. Often it’s faster and easier to edit incorrect autogenerated code than to write it from scratch, especially when you work with many languages with their own idioms. Reviewing code is part of the process whether I’m writing from scratch or using autosuggestions.

So your answer to “what makes you optimistic about the future?” is the apocalypse? Reminds me of pets by porno for pyros

How did you find an editor? Do you pay them or do they work pro bono?

Does anyone know of this is the case? Are the earnings from prison labor taxed?

The author provides some detail on the process here https://www.patreon.com/posts/i-used-ai-to-66518281

What are the biggest benefits over Alfred?

Don’t try to catch a falling knife.

That’s a good question. Does anyone have an answer or argument for why this wouldn’t be disturbing to cats and dogs?