wanderingbit

I’m confused why this isn’t talked about more often.

As a parent, most of my time that I spend thinking about money is thinking about how to pay for college. It is the second largest purchase im going to make next to my house, and it is the financial decision that sucks up the most thought.

If my kids have a 1/3 chance of interacting with someone that doesn’t think college is worth it, then that impacts how I think about whether to orient the next 20 years of savings around saving for my kids college.

If that person interacting with my kids is their boss, or person tasked with hiring them, then this _greatly_ impacts my decision to pay my kids’ college.

Note: I don’t mean to imply “I’m deciding between whether my kids have a dedicated time period where they live on their own and just study”. I’m specifically talking about paying for an education experience that will result in a credential (a degree).

Maybe I’m greatly impacted by Bryan Caplan’s “The Case Against Education” but some part of me believes the main value of a college education is that other people think a college education is valuable.

This seems to be getting less important in my lifetime. It’s hard to tell where the general sentiment is gonna be 20 years from now.

:(

I set this up on my machine and tested it with the following question:

How can I build a bow drill so I can start a fire?

It got it partially right: spindle, bow with string tied the correct way, and fire board. But then it flubbed by saying you put the spindle between both hands and move your hands back and forth to rotate the spindle and create the necessary friction… this totally ignores using the bow to rotate the spindle! It also mentioned “fanning the sparks” but there is absolutely no way your bow drill is going to create sparks.

I asked a frontier model to check the survival LLM’s conversation trace and it pointed out the errors.

I didn’t see obvious errors in the other 3 prompts I used, but this one seemed pretty egregious for a survival LLM.

I thank the creator for making this, but take it with a grain of salt.

This would have saved customers at a previous company of mine weeks of developer time, thank you!

Can you include a mention of the signing key algorithm (ed25519) higher up in the README? That was important information I had to scroll a long ways to find.

How does this compare to using something like exe.dev where you get remote isolated sandboxes over SSH, and assume you add in an http credential injection reverse proxy (bring your own)?

I'm building something that will give me peace of mind that my agents will (whether maliciously or stupidly) not negatively read/write the personal data on my machine. To me, exe.dev (with possibly a reverse proxy to preclude the agent from ever having access to my API tokens) is sufficient.

We are trying to design our kitchen for a renovation and I can see how these 3 constraints would be useful for us to do for something more about design than software.

I’m gonna go do these…

This has already helped me find extremely useful writings on a niche topic of mine. Thank you!

@sandGorgon8, thank you.

I was just telling some ex coworker friends that there was a great need for a compose frontend to more powerful infra backends, and this feels like the answer.

Once I get working on it I’ll try to add health check support. That is crucial for a lot of what we’re working on.

> And they didn't even bother to test the most important thing. Were the LLM evaluations even accurate!

This is not true; the professor and the TAs graded every student submission. See this paragraph from the article:

(Just in case you are wondering, I graded all exams myself and I asked the TA to also grade the exams; we mostly agreed with the LLM grades, and I aligned mostly with the softie Gemini. However, when examining the cases when my grades disagreed with the council, I found that the council was more consistent across students and I often thought that the council graded more strictly but more fairly.)

This is top 10 for greatest HN deep dives. I learned something new almost every sentence, and could not complete it on my first attempt.

There are a couple big problems with this type of digital and decentralized type of authentication (I say this as a long time cryptocurrency professional who wants this to succeed):

1. backups and account recovery: We’re working with humans here. They will lose their keys in great numbers, sometimes into the hands of malicious actors. How do users then recover their credentials in a quick and reliable manner?

2. Fragmentation: let’s be optimistic and say digital credentials for drivers licenses are given out by _only_ 50 entities (one per State). Assuming we don’t have a single federal format for them (read: politically infeasible national id) how does facebook, let alone some rando startup, handle parsing and authenticating all these different credential formats? Oh and they can change at any time, due to some rando political issue in the given state.

OP, you clearly know all this, so I’m just reminding you as someone down in the identity trenches.

You seem like a knowledgeable crypto user. Can you help integrate lightning payments with x402? Then we'll be rid of this Base/corporate-tech nonsense.

I’ve been looking for something like this. The problem is, when to find the time to do it.

This isn’t enough information about the study to tease out cause and effect. There may be a third confounding variable that positively impacts both entrepreneurship growth and Starbucks growth.

For instance, what if Starbucks only decides to move into neighborhoods that have reached a certain level of economic growth (ie number of households, number of business, etc…)? Neighborhood economic growth would likely attract entrepreneurs as well, and we wouldn’t be able to conclude that Starbucks had anything to do with entrepreneurship growth.

Said a different way, would adding Starbucks in the middle of the Atacama desert grow Peruvian entrepreneurs? I mean come on it’d be the only third space around!

I can’t read the full paper because I don’t have a subscription, but the fact that they don’t call this out in the abstract makes me doubt it’s a meaningful conclusion.

The author mentions putting 20% down and getting a mortgage. Even with insurance, for a 10 year mortgage 29k is 430 per month. Or you live with your parents and save on $1500 rent for 2 years, bam you can buy it with cash.