David Shapiro ⏩ (@daveshapi) 's Twitter Profile
David Shapiro ⏩

@daveshapi

AI Maximalist, Psychedelics Advocacy, Post-Labor Economics

20+ years in tech, 10+ years in AI/ML

ID: 1795781228399869952

linkhttps://linktr.ee/daveshap calendar_today29-05-2024 11:36:37

13,13K Tweet

36,36K Followers

851 Following

David Shapiro ⏩ (@daveshapi) 's Twitter Profile Photo

When I came across this trend in my research I had to double check. Triple check. The causes are many: Some men are getting left behind. Others are disabled or addicted. A few leave voluntarily. By and large, this is a secular trend. Automation is a strong contributing

David Shapiro ⏩ (@daveshapi) 's Twitter Profile Photo

I think that the UK should be celebrating foreign investment, considering the current state of civil and economic affairs. But if this attitude is common, that explains a few things.

David Shapiro ⏩ (@daveshapi) 's Twitter Profile Photo

This is a MAJOR step change in the future if autonomous AI. It's difficult to overstate how much this will have ripple effects for decades to come.

David Shapiro ⏩ (@daveshapi) 's Twitter Profile Photo

If your wealth is tied up in stocks, then "the economy" you care about is the stock market If your wealth is primarily tied up in your labor, then "the economy" you care about is the labor market These two markets often have antagonistic interests.

David Shapiro ⏩ (@daveshapi) 's Twitter Profile Photo

AI haters often say it's a nothing burger but when you have to throttle your users because you don't have enough supply that generally is a good thing for a company. It's odd how many people gloss over this fact.

David Shapiro ⏩ (@daveshapi) 's Twitter Profile Photo

A while ago I said that I could foresee a future in which is would be illegal for a human to drive on many roads. If we got to a point where machine drivers are that much safer than humans, it would be unethical to endanger people. With that being said, I suspect the market

David Shapiro ⏩ (@daveshapi) 's Twitter Profile Photo

The reason it comes off as "subtle" is because it was rhetorically refined on forums for the last 10 decades. The "test" and "experimentation" was "does this blob of words convince ordinary internet denizens?" That was the loss function. It had nothing to do with actual AI,

The reason it comes off as "subtle" is because it was rhetorically refined on forums for the last 10 decades. 

The "test" and "experimentation" was "does this blob of words convince ordinary internet denizens?" 

That was the loss function. It had nothing to do with actual AI,
David Shapiro ⏩ (@daveshapi) 's Twitter Profile Photo

I randomly got charged $25 by anthropic. I thought I canceled my subscription months ago. I went to investigate It was the API platform. My $25 of credits from a year ago expired (which I hadn't used) and auto-refill was enabled. So now auto-refill is disabled.

David Shapiro ⏩ (@daveshapi) 's Twitter Profile Photo

One of the least credible aspects of the Star Trek world is that they acknowledge, in universe, that automation could do all their jobs, but they insist on doing shoddy, slapdash, human-centric jobs (and don't even really rely on machines for help) on principle.

David Shapiro ⏩ (@daveshapi) 's Twitter Profile Photo

TL;DR The paper introduces “K-Dense,” a guided, multi-agent AI workflow that helped the authors build a transcriptomic “aging clock” that reports not only a biological-age estimate but also a calibrated confidence interval. Using 57,584 RNA-seq samples from 28 tissues (ARCHS4),

David Shapiro ⏩ (@daveshapi) 's Twitter Profile Photo

OpenAI is having an interesting day. I would anticipate an AI Explained video any minute now... Between this and their ICPC announcement, it should be interesting.

David Shapiro ⏩ (@daveshapi) 's Twitter Profile Photo

Okay, here's the simple metaphor for how OpenAI cut "scheming" rates from 13% to 0.3% in some models: You know how in math class you'd get graded as much (or more) on the METHOD you used? SHOW YOUR WORK. They want to see you following their first principles, their reasoning,

David Shapiro ⏩ (@daveshapi) 's Twitter Profile Photo

I am once again reminding you that you are not excited enough about this graph. It looks like a simple exponential, right? But it's on a logarithmic graph...

I am once again reminding you that you are not excited enough about this graph. 

It looks like a simple exponential, right? 

But it's on a logarithmic graph...