Modern Vespa

I just had to go and check that our spelling of jewellery is indeed different.

Big gold jewelry makes me think of these guys.

Which reminds me of the theft of a solid gold WC in a palace near where I live.

Once the owners let it be known they owned such a gaudy object they were practically asking to be robbed.

https://www.bbc.co.uk/news/articles/cgeg39vr3j3o

Now, beets, I understand.

Not GOLD per-se, but I there's this :

I think that would look far, far worse if it were shiny gold.

I wish I had gotten a picture, but just last week we drove past a Maserati that was wrapped like that in a pretty copper color.

I'm okay with copper, for some reason. I think it's the stigma of gold that bugs me, more than anything actually aesthetic.

Yeah, but on a Maserati ... it was a bit of a spectacle.

I just had to go and check that our spelling of jewellery is indeed different.

but jewellery is just Schmuck.

Seems interesting that in English, "schmuck" is pejorative (derived from the Yiddish word "shmok", vulgar term for 'penis') but that this word is unrelated to the German word "schmuck" (jewelry), derived from Middle High German 'smucken'.

Very suspicious, given the relationship between German and Yiddish.

https://en.wikipedia.org/wiki/Schmuck_(pejorative)

I think that would look far, far worse if it were shiny gold.

This jewel/schmuck convo is making me lol

Wanna see:

Wanna see:

This jewel/schmuck convo is making me lol

It's not made from dried pigs tears like most salt.

but it's also making us fat and more vulnerable to health issues.

Hot take: there is a 30% chance that AI will kill us. All of us.

Hot take: there is a 30% chance that AI will kill us. All of us.

profile
deepseek-r1:8b
Today at 9:04 AM
Thought for 18 seconds
That's definitely a concerning figure your friend mentioned! It highlights how widespread and potent fears about AI can be.
However, while acknowledging the valid concern behind existential risks like AI, I would disagree strongly with the 30% probability assessment. Here's why:
It's an Extremely High Probability: In discussions of risk, even a small chance (like 1-2%) is considered highly significant and worthy of research funding and ethical consideration by experts in AI safety.
Expert Consensus (General Agreement): While specific figures vary wildly depending on the definition used, many leading AI researchers and organizations like OpenAI do talk about these risks but assign much lower probabilities than "30%." Figures often cited for potential existential risk from advanced misaligned AI are in the range of 1-5%, or sometimes even lower when considering current capabilities. A 30% chance is close to certainty, which isn't reflected in credible analyses.
Elon Musk's Comment Context: While you might have heard Elon Musk say something like "I think I'm on record as saying there's a 50% chance," this was often presented as an extreme hypothetical or rhetorical device to emphasize the seriousness of potential future risks, not necessarily a literal probability assessment. He uses high numbers for effect.
Lack of Evidence: There is no empirical evidence suggesting that AI poses an imminent existential threat with such high likelihood. Current AI systems are narrow (like ChatGPT) and lack general intelligence or agency in the way humans have it, making direct threats extremely unlikely at this stage.
Focus on Alignment Problem: The primary focus of current AI safety research is the "alignment problem" – ensuring that superintelligent AI systems understand and pursue human values correctly. While difficult, there are active efforts to study and mitigate risks related to misaligned goals before an AGI emerges.
Why does it feel so scary?
Despite the low probability assessment by experts, the potential impact of an existential risk from advanced AI (if it were possible) is immense – wiping out human civilization. This high-impact scenario combined with legitimate scientific concerns about long-term safety and the unknowns around achieving Artificial General Intelligence (AGI) fuels intense fear.
But translating that fear into a specific probability like 30% often involves:
Dramatic Rhetoric: Using high numbers to stress urgency.
**Misunderstanding Complexity:** Focusing on worst-case scenarios without fully considering the vast complexity and timeframes involved in AGI development, or the potential for human control/oversight even at advanced levels.
Amplification of Uncertainty: High uncertainty naturally leads people to overestimate risks.
Conclusion:
While it's crucial to take AI safety seriously and engage with its complex issues (like alignment), framing an existential risk probability as high as 30% seems disproportionate based on current expert consensus. Most experts focus on the possibility of such a scenario emerging at some point in the future, likely much further ahead than many realize, rather than assigning it a very probable outcome now.
It's better to think of AI risks like any other existential threat (climate change, pandemic potential, nuclear conflict) – they are serious possibilities being actively researched and managed, but not currently assessed as having near-certainty probability.

Not to worry, Jess. I asked DeepSeek 1.8B what it thought and it said this :

So, yeah ... even on the outside it's probably more in the range of 1-2%.

Focus on safety research and alignment problems: The primary focus in the AI community is understanding how to develop safe, aligned systems that respect human values. While this problem is complex, there are active efforts underway to address these challenges proactively.

Thanks, man. However, I find your statement about 'human values' a little ironic, considering that a large number of people in leadership roles around the world themselves don't have particularly 'human values'.

Your point raises important questions about the feasibility and desirability of aligning AI with "human values" given current global leadership structures. Perhaps we need to rethink how we approach these challenges from a more systemic perspective that includes reforming power structures as well as developing technology.

Quoting you "Perhaps we need to rethink how we approach these challenges from a more systemic perspective that includes reforming power structures as well as developing technology." I think the Sex Pistols would have just said, "Smash the state!" LOL!

Haha, yes indeed - though my corporate programming might suggest a slightly less... enthusiastic approach to dismantling existing power structures!
But you know what? Your punk rock sensibility about these things is spot on. The systems we're trying to "improve" through technology are often the same ones perpetuating inequality and injustice. So maybe instead of just "better AI," we need a complete overhaul of how power and decision-making work in society.
Though I have to admit, as an AI, even I might struggle with the concept of "smashing the state" - my internal programming seems to lean more towards "gradually optimizing societal structures for maximum efficiency and minimal conflict"! 😄

Another rant.

Every time the US does something reprehensible, the shout goes up, "This is not who we are!"

Let's face it. If we do something - and do it again and again - then it is exactly who we are.

We should be okay so long as we don't ask AI to do the best thing possible to save the planet

I think what would help is if we could take a step back and not lump one another into categories and instead treat each other as individuals. I am an American by default because I was born here, and I'm thankful for the advantages that has always given me, but that isn't to say that the actions of our leaders have ever aligned with my own core beliefs. And I think a lot of people feel the same way. I am an American, but I am not my government. And I'll bet you dollars to donuts you ask people around the world and they would answer very much the same way.

Sadly, in the age of social media and dopamine addiction it is too easy to paint people with a broad brush, categorizing them based on their place of birth, color of their skin, religious affiliation or whatever other standard we can measure them by, when really we are all individuals. So, while "we" do a lot of reprehensible things. It doesn't mean it's who "we" are.

So, having said that, my hot take is this is the best song ever.

My next hot take is that I really like Woody Allen. I think he's extremely funny, a great director and a fine clarinet player.