Promethease (1) is a great site for doing this. You upload a genome (which they promise to delete), and download an HTML interactive report. The report spans thousands of studies, and it will tell you about any rare genetic disorders or interesting findings.

The cost is about $18, one time, iirc.

(1) https://promethease.com/

Do not stop at Promethease! I have been doing this for myself and others for the last 10 years. Promethease is great but misses several SNPs that are included in the raw data that may have an impact. Some other helpful websites:

https://www.ncbi.nlm.nih.gov/snp/

https://www.uniprot.org/

https://www.snp-nexus.org/v4/

I have used Nebula. Did a group buy of the kits with several friends to obfuscate shipping address. Signed up via a protonmail address, and accessed the site only through VPN. Then deleted my data after downloading all of it.

I believe it would be very difficult to tie my identity to my genome because of these steps.

> I believe it would be very difficult to tie my identity to my genome because of these steps.

That probably depends on the closest relatively of yours that they have in their database.

If you only have one sibling and they're in the database, you're probably boned no matter the precautions you take.

Yep. Standard clickbait formula, stir up emotions, add no new information. So easy an AI could do it.

Which is a bit surprising, given that RL is the most likely path to the scary AI future everyone loves to blog about.

Doesn't surprise me. Maybe you've heard of low temperature heat pumps, but not the refrigerants that enabled that technology. You've certainly heard about solar panels but you probably don't know what metallurgical and manufacturing advances made them cheap enough for many people to buy.

Why should someone learn technical jargon when what there interested in is what effects the technology will have on society?

An article criticizing the US with a .cn domain. Chinese propaganda.

"Average % Revenue growth in pre-seed companies" is a very noisy metric.

Example:

Suppose you had one company doing $10 in sales a year, now it's doing $100 per year. 900% growth!

When you average 900% with the rest, it will skew the results wildly.

Yeah, it is even possible to do this sort of analysis in domains with exponentially-distributed outcomes?

Yes, but it will be incredibly misleading. You'd want to look at the distribution of outcomes to get a sense of what's really happening. The startup I work for is remote first but has some offices so we wouldn't count as fully remote even though we mostly are.

All these AIs are doing is sampling from some continuous space. Every answer is a "hallucination".

AIs are useful because many (most?) of those hallucinations happen to be useful.

Even the human vision system is hallucinating all the time. The difference is that when we're awake we have another filter that separates the unreal hallucinations from the ones that are much more likely to match reality.

I experienced this very memorably coming back from vacation where there were little lizards seen while walking to and from the beach. When I got back home to my cold climate, I 100% saw a lizard skit across the sidewalk--it was a leaf.

That's using "hallucination" the way humans use it.

My understanding was that the term as used in the field is fairly well defined as "producing a confident answer that is not backed up / justified by training data". It has nothing to do with sentience or humanness.

An AI model that can only repeat its training data would be no better than normal text search.

AI models are useful precisely because they can interpolate - make guesses that incorporate information from many examples.

And this interpolation is called "hallucination" only when it happens to be a wrong answer?

It's not a hallucination. When a parrot speaks it isn't sentient. LLM's are essentially doing parrot math with much larger data sets, that is it. The fact that it sufficiently mimics what you're expecting is not a testament to its intellect.

>When a parrot speaks it isn't sentient

Parrots are certainly sentient. You could maybe say parrots aren't sapient, but frankly I think we don't know enough about sapience or what it takes to make human-level intelligence to make that claim either.

Selling it as "free up the parents to do other things" is not going to fly. But enhancing the baby's experience or learning is fine.

An AI-powered rocker that plays soothing music till the baby sleeps would be great. People already use rockers, it's very accepted.

I would never have trusted a computer to be in charge of my infants. Even trusting another human adult is pretty tough already.

The Control Problem is scary enough already. What's the Paperclip Galaxy equivalent for a robot nanny?

If you rented out that house, what would you charge? Would you charge 3k, and lose 2k / month?

That doesn't make sense; so how are landlords surviving in your market then?

They probably bought it when it was cheaper or own outright. Sometimes people will rent at a loss hoping for appreciation, though 2k is a bit steep.

No opinion on whether the US did it.

But Hersh's essay indicates there were last-minute changes to the detonation method. That adds risk, which could have manifested as failure of some detonations.

Then the explosive charges are still there, and Danemark and Germany can recover them.