ipython

Sounds like a crude cost/benefit analysis to me :-p

It used to be worse. I operate a vpn for my extended family, some of whom are deployed overseas at any time.

They would google for things and suddenly my ads would show nightclubs near them (thousands of miles from me) and google’s default language would even change to the country they currently reside in. Just because the outgoing ip is shared across both users.

It’s actually gotten “better” but one could argue maybe they’re able to perform more precise targeting instead of throwing away signal.

> This is also a signal what what the “real cost” of these services will be once the VC subsidies dry up.

Sooooo true. Waiting for AI’s “uber” moment aka when uber suddenly needed to actually turn a profit and overnight drivers were paid less while prices rose 3x.

I find list and dict comprehensions are a lot less error prone and more robust than the “manual” alternatives.

I would say unless you have a good reason to do so, features such as meta classes or monkey patching would be top of list to avoid in shared codebases.

The backend can. But what’s exposed to customers will be a very very small subset of that capability. Hence why only the csrs can perform that function.

The business undoubtedly did a crude cost/benefit analysis where the cost to expose and maintain that public interface vastly outstrips the cost for the few people that have to call in and change their name.

I think we are starting to see that under the current administration, yes. But the question was - does it rise to the level that Chinese have? I argue no.

I also worry about some sort of policy where a site “must” carry content. That has its own risks as well, which should be obvious.

Totally not the actions of a shady company with cowboy culture built in from the top: https://www.reuters.com/technology/tesla-workers-shared-sens...

What is “deboosting”? If I own a social media site wouldn’t you argue that I should be able to control what content I “boost” in the first place? Am I forced to compulsory “boost” your content? In which case wouldn’t that mean I am “censoring” the PoV of the social media company?

Can you share the topic contents that are subject to your accusations? If you’re able to share them here without fear of government reprisal then I would argue that’s not censorship.

Really? Who do you think douyin hires to perform content moderation for example?

https://dl.acm.org/doi/10.1145/3706598.3714013

Also, don’t forget to obtain your icp license from the government before establishing a website in the mainland!

I am torn. I see this argument and intellectually agree with it (that interfaces need to be more explicit). However it seems that every time there is a choice between “better” design and “good enough”, the “good enough” wins handily.

Multics vs Unix, xml based soap vs json based rest apis, xhtml’s failure, javascript itself, … I could keep going on.

So I’ve resigned myself to admitting that we are doomed to reimplement the “good enough” every time, and continue to apply bandaid after bandaid to gradually fix problems after we rediscover them, slowly.

They don’t say what the kWh usage is, just that the electricity cost in $$ is over $1000 on the highest month. For a unit surrounded by what should be other conditioned spaces, that’s insane to me.

A quick web search indicates that nyc $/kwh is about 31c. So that’s 3225kwh in one month! My standalone house plus pool pump, dual zone ac, and ev charger doesn’t even come close. Clearly there is a major insulation issue which is the root cause and everything else is just trying to put bandaids on an arterial bleed.

I would demand an itemized bill from that urgent care and play hard ball with that provider. That's absurd, even by US standards. I've had ER visits that total less than that. My son had a cast that fell off while on travel in Florida, and we stopped by an ER (only place that was open and could fit a new cast). Total cost out the door _at the ER_ was less than $500, which while expensive, is nowhere near what you're talking about.

The first factor is access to your email. The second factor is…?

So I’ve heard of this model collapse theory. But I’ve also heard of model providers who are intentionally training with synthetically generated data (as a result of insufficient “real” data).

So I’m curious where the line is? Are there phases in the training/continued pre training/alignment/rlhf pipeline where synthetic data isn’t just harmless but actually beneficial? Is it a question of quantity or a question of how much novelty is in the training data?

Just wanted to remind everyone, journalism like this is partially funded by the Corporation for Public Broadcasting, which was just defunded by this administration this week. See their "Funders" section here: https://features.apmreports.org/about.html.

+ HOA fee (condo HOA fees can easily add up to several hundred $$ per month)

Have you seen pics of Houston? It can be infeasible to walk a mile if it’s split by a multi lane highway etc.

Fantastic point. Hard to scale stealing data from files stored in a physical location. Easy to scale exfiltrating data via credential stuffing.

What voip service allows free calling to/from pstn? Without an app filled with ads?

So glad to see a renewed emphasis on proper wired infrastructure. It seems the "big boys" (Verizon, T-Mobile, etc) are heavily pushing wireless and not building out new wired areas, I assume because it's less capital intensive.

Hell if there's a way to invest in Prime-One, these guys seem to have their stuff together...

This article is based upon an interview from NYT: https://www.nytimes.com/2025/06/26/opinion/peter-thiel-antic....

That NYT interview with Ross Douthat is just a lame attempt by Thiel to rehabilitate his image. From the interview, regarding how we as a society entered into a state of technological stagnation, and his subsequent support of Trump in 2016 as a way to "redirect the Titanic from the iceberg it was heading to, or whatever the metaphor is, to really change course as a society":

"I didn’t have great expectations about what Trump would do in a positive way, but I thought at least, for the first time in 100 years, we had a Republican who was not giving us this syrupy Bush nonsense. It was not the same as progress, but we could at least have a conversation. In retrospect, this was a preposterous fantasy."

Really? You, arguably the most influential and successful inventor of the modern age, couldn't foresee that Trump is just going to leverage his populism to further cement the power structures that brought him into power? How was that ever going to lead to societal level change in a way that would bring about an explosion of science and technological innovation?

Nah, I'd believe that Thiel (correctly) sensed that Trump was a vessel through which he could accumulate much more power and wealth, due to Trump's brazen cronyism and corruption.

So now Thiel needs to somehow distance himself from Musk, Trump, and everyone else to try and recover whatever dignity he can. The only good news I can see is that, if he truly has any sort of crystal ball, he senses the winds starting to shift and the political power of MAGA waning.

I loved that game! Used to play it on my palmpilot and compete with my workmates to see how much $$ we could make.

Agreed! I guess I don't understand as I have seen five year olds running lemonade stands with more business sense than this LLM.

More to the point, if you steal the book and never even read it, you are still guilty of a crime.

Exactly. I feel like the AI companies are intentionally moving the goal posts- regardless of whether the resulting generated content is the same as the original, they still committed the crime of downloading and using the original copyright content in the first place!

After all they wouldn’t have used that content unless it provided some utility over not using it…

I just had this argument with a state wide government website. I have to log in to this site maybe once per year to update contact information and update a few fields. Unfortunately, that site silently deactivates your account automatically every 90 days. So I'm forced to change the password literally every time I log into the dumb thing.

They refused to establish MFA or passkeys - and instead insist that "NIST is the minimum recommendation for cybersecurity... and we take cybersecurity very seriously... to ensure the safety and security of the citizens... therefore we will not change our policy on mandatory account lockouts or password change requirements."

Azure OpenAI absolutely got paid.

I'm not sure exactly where you get 12 lines of code from. There are at least a hundred lines of LLM-related "prompt engineering" alone: see https://github.com/slavingia/va/blob/35e3ff1b9e0eb1c8aaaebf3..., for example.

Regardless, number of lines is not an accurate metric of how "essential" a component is to the functioning of the overall system. How about this, try removing the LLM-related code from the script... what functionality is left, exactly?

There’s a whole sect who believes the opposite- that you are more spiritual and blessed by God the wealthier you are. Somehow being material wealthy is now a signal of your spirituality. :shrug:

https://en.wikipedia.org/wiki/Prosperity_theology

Ok. From the party of “states rights” that’s a bit hypocritical of them. I mean- they applauded Dodds which basically did the exact opposite of this- forcing states to regulate abortion rather than a uniform federal standard.