MPSimmons

See, now I'm thinking of mayonnaise injectors to make instance turkey salad

It's unfortunate that this happened, but this will help drive better engineering decisions in the future for everyone. Glad everyone is mostly okay from this!

>The test correctly ruled out cancer in over 99% of those who tested negative.

Forgive me if I'm wrong, but isn't this the textbook example of understanding false negatives in testing people at scale?

Hanlon's Razor

Maybe by that time ARM linux on this platform will be excellent and we can migrate to it for old gear. I still have a 2011 MBP running Linux on my electronics workbench and it is just fine.

Doesn't the click cause the browser to "go" on mouseUp? Selecting is clicking then dragging - this seems like a clear enough difference to me that I've never had problems. In fact, sometimes I will click, and think, "oh I actually don't want to go there" and will drag off the link and release the mouse button and it doesn't take me there.

I would hope fervently that HackerNews would be subject to selection bias and would be an exception, but who knows.

I caught that too. What's a "pre-idea" individual? Someone who... wants the vague _idea_ of a company?

It wasn't just about the subscribers. Blogs were incredibly useful as long-form documentation of unique problems and experiences, too. I used to write in my blog (Standalone SysAdmin), and it had over a million hits per year a few times, and the majority of those hits which didn't come from going viral about ashtrays in airplane bathrooms were people searching for the same problems I'd encountered and written about.

The crux is how big the L is in the local LLMs. Depending on what it's used for, you can actually get really good performance on topically trained models when leveraged for their specific purpose.

I actually thought you were trying to introduce training data to make AI artificially fail on Christmas

Is there no Argo plugin for your secret store? In a previous life, we used Argo Vault Plugin to good effect.

I hate to be like, "Let's use AI for this" on everything, but this actually does seem kind of like a good thing for AI to do, rather than subject a human to it.

Just because _someone_ is going to do something evil doesn't mean it has to be _you_. Every person has a choice, and he chose to do evil.

I feel like you're making a bad joke. Did you drop your /s?

I had the thought the other day that one of the most valuable things a human-driven website could offer would be a webring linking to other human-driven websites

Immediately did ctrl-f "modeline" and was pleasantly surprised

Is it likely that the Executive Branch would try to exert control over it to remove "inconvenient" data?

Agree with your experiences. I've also found that if I build a lightweight skeleton of the structure of the program, it does a much better job. Also, ensuring that it does a full fledged planning/non-executing step before starting to change things leads to good results.

I have been using Cline in VSCode, and I've been enjoying it a lot.

I believe it has to be kept at refrigerated temperatures over its relatively short shelf-life (1 year, I believe).

>a high enough score is a threshold to claim AGI

I'm pretty sure he said that AGI would achieve a high score, not that a high score was indicative of AGI

I don't think we actually even have a good definition of "This is what AGI is, and here are the stationary goal posts that, when these thresholds are met, then we will have AGI".

If you judged human intelligence by our AI standards, then would humans even pass as Natural General Intelligence? Human intelligence tests are constantly changing, being invalidated, and rerolled as well.

I maintain that today's modern LLMs would pass sufficiently for AGI and is also very close to passing a Turing Test, if measured in 1950 when the test was proposed.

Yep. Lots of examples on YouTube: https://www.youtube.com/watch?v=xV2gsWHhDhY

Agreed. It's basically lossy compression for everything it's ever read. And the quantization impacts the lossiness, but since a lot of text is super fluffy, we tend not to notice as much as we would when we, say, listen to music that has been compressed in a lossy way.

Neither getting 40MW of power nor removing 40MW of heat are easy.

The ISS makes almost 250KW in full light, so you would need approximately 160 times the solar footprint of the ISS for that datacenter.

The ISS dissipates that heat using pumps to move ammonia in pipes out to a radiator that is a bit over 42m^2. Assuming the same level of efficiency, that's over 6km^2 of heat dissipation that needs empty space to dissipate to.

That's a lot.

A weeble-wobble

I've always seen a D1 as a bingo ball...

Likely longer as this was the only test site for Starship and there will hopefully be an FAA investigation.

In spaceflight, an anomaly is an anomalous outcome compared to what was expected. The severity of the anomaly varies. Typically, if the world outside of the organization hears about an anomaly, it was severe enough to cause a Loss of Mission (LOM) or Loss of Vehicle (LOV). Internally, when things behave anomalously, they're referred to as off-nominal, and are subject to internal investigations to determine the cause.

This is a _very_ off-nominal outcome and the investigation will absolutely involve outside organizations and halting the program during the investigation until the investigation completes with a sufficient determination of faults and accompanying remediation plans.