For your first question, I think the average person would benefit from a simple digital currency that let’s them exchange “cash” without having to jump through a bunch of hoops. Venmo, Zelle, etc. are all proof that normal people want easy ways to pay each other.
As for your second point, I’m not sure I follow. But I assume you’re implying that crypto is better because it isn’t tied to the state?
So I’ve been reading into stable coins a lot, because I don’t understand why anyone would care about them. And what I’ve come to realize is they are a benefit to two groups:
At the end of the day, it feels like a “true” digital currency would be the better solution, but everyones jumping on Stablecoins because they’re here now and less regulated.
I think there is potential in a more cash-like digital currency, but Stablecoins seem ripe to break in some unforseen way, especially given the current administration.
Edit: Edited to fix formatting.
We definitely haven’t cracked AGI, that’s without a doubt.
But yeah, LLMs are big (I’d say really Transformers were the breakthrough). My point though was that Deep Learning is the underlying technology driving all of this and we certainly haven’t run out of ideas in that space even if LLMs may be hitting a dead end.
That’s the issue, AI right now means LLMs not deep learning/ML.
The Deep Learning/ML stuff will keep chugging along.
I think we may see an agentic AI winter, but there are so many other applications and systems that can benefit from deep learning still.
Yeah, but that doesn’t help if you can’t make apps that support the hosted services. Google is trying to have complete ownership of what runs on your phone.
There are laws about how to handle PII and potential criminal charges based on things like the Privacy Act. Meaning there are additional requirements above and beyond how people normally store data on a system.
More requirements = More chances to mess up
I think the line “how does handling PII make it easier to mess things up” just about sums things up for me.
While AWS/Azure do make the initial configs rather fool proof, that falls apart the moment you start configuring them for actual use. It’s also especially easy to mess things up when handling PII, at the SSA level it’s probably something that DOGE staff don’t have experience with.
As for vaccines. Largely through that out there cause it seemed like obvious bait for you, but I don’t think a single slogan “my choice my body” really encapsulates the arguments around abortion
In cyber security you may never know if a bad actor got access to your systems/data. The issue with not following good security practices is that you increase the risk of this happening.
Its like saying we should stop mandating vaccines cause the diseases aren’t around anymore. When you let down your defenses you end up with outbreaks that shouldn’t have happened and are harder to control.
The SSA stores a lot of sensitive data. Normally with sensitive data you want to be very careful with who can access it and how.
What is potentially worrisome in this situation is it seems like the SSA is taking on the “move fast and break things” attitude of Silicon Valley.
More technically, most government agencies use AWS and Azure (cloud providers) to host data. So spinning up a new server isn’t inherently bad. However, creating a new server that is secure and has the correct access controls (user permissions regarding who can see/change content) can be challenging. The whistle blower believes they are not doing this right, and it sounds like the head of the SSA isn’t disagreeing, just saying he thinks the risk is worth it.
Given it’s the government it’s most likely AWS or Azure. That really isn’t inherently bad, it’s more the attitude of “move fast and break things” doesn’t necessarily work for secure systems with sensitive data.
If you read the article, the current head of the SSA acknowledges they did set up the system being discussed and that he’s accepted the increased risk of the implementation as there is a “business need”.
I agree that “random server” is a bad choice of words, but do want to add additional information context as the concern isn’t necessarily unwarranted. Another qoute from the article:
“I have determined the business need is higher than the security risk associated with this implementation and I accept all risks,” wrote Aram Moghaddassi, who worked at two of Mr. Musk’s companies, X and Neuralink, before becoming Social Security’s chief information officer, in a July 15 memo.
Its also sounds like they did spin up a new database with limited security/oversight to “move” faster. Why that’s worrisome is they aren’t denying there is a risk or lack of security, they are just saying it’s justified.
I think it’s a mixture of that and the fact that when OpenAI saw that throwing more data drastically improved the models, they thought they would continue to see jumps like that.
However, we now know that bad benchmarks were misleading how steep the improvements were, and much like with autonomous vehicles, solving 90% of the problem is still a farcry away from 100%.