Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
mythhabit
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
8 ms
·
61.
▲
by
mythhabit
3y ago
Both France and Germany have nuclear plants with load following capabilities to the tune of 5% change per minute. That is 50-100MW change every minute for a single powerplant. It might not be peak-peak, but it is certainly going to handle t
62.
▲
by
mythhabit
3y ago
The majority of "difficulty" in software development is not writing the code. It's specification. And while that may or may not be solved by LLM or other AI tech, we're so far off that it's not even a thing right no
63.
▲
by
mythhabit
3y ago
If she gets to sleep through the night, she gets to wake early and take care of the child while you get a few hours uninterrupted sleep. Sleep is as important for your mental health and longevity as exercise is for your physical health and
64.
▲
by
mythhabit
3y ago
“There’s been a systematic change where employees feel the employer is extremely lucky to have them, as opposed to the other way around,” he said. “We’ve got to kill that attitude and that has to come through hurting the economy.” Or, maybe
65.
▲
by
mythhabit
3y ago
If my bank cold calls me, I can say "I just need to verify the legitimacy of your call, so send me your direct number in the online bank app, and I'll call you". It works every time, but it also works because all the employee
66.
▲
by
mythhabit
3y ago
We have strong standards - but if there is only one implementation, that quickly becomes the standard no matter the text in the actual formalised standard.
67.
▲
by
mythhabit
3y ago
In Denmark, the difference in interest is 1.5-2.5% between a fixed interest loan and a variable one. Historically it has always been a good idea with a variable interest loan, but that changed after covid.
68.
▲
by
mythhabit
3y ago
And whatever Google decides they want to implement or support in Chromium will be the web default, even just how they interpret ambiguities or undefined behaviours - standards be damned. We have tried this before and it was not pretty. That
69.
▲
by
mythhabit
3y ago
In Denmark, all lenders must be able to take a 30 year fixed interest mortgage. If they can do that, and still have additional room in their finances, they can be allowed to take home a more risky loan, all the way down to interest adjust e
70.
▲
by
mythhabit
3y ago
I believe that there is merit to the theory. If you bought a house 5-7 years ago, the interest and taxes is half of what it cost to rent a smaller house in the same neighbourhood. The high rent is not driven by demand, nor by cost. It'
71.
▲
by
mythhabit
3y ago
You should never put people in jail for using any controlled substance. That is mostly an american thing. For people using it, just fine them and be done with it. Go after the people producing, distributing and selling.
72.
▲
by
mythhabit
3y ago
The research is old and in its infancy, but I believe there is a pretty clear evidence that link underlying predisposition to certain mental illnesses and a psychotic episode when using psychedelics. I believe it remains to be determined wh
73.
▲
by
mythhabit
3y ago
I think the main distinction is that Apple claims to have a secure phone, but not an unhackable phone. A secure vault is hard to get into, but not impossible. Should they have done something about this? I believe so, but they are not market
74.
▲
by
mythhabit
3y ago
That does not change anything, asymmetric or symmetric, the same principles apply. E2E means that only the recipients has the ability to decrypt the messages, so far we seem to agree. As I said, and you verified, for your scheme to work, th
75.
▲
by
mythhabit
3y ago
Centrally managed E2EE is not end to end if the server can decrypt anything. The definition means that the keys to decrypt only exists at either end.
76.
▲
by
mythhabit
3y ago
If a company have the means to decrypt a particular users data, they have the ability to decrypt all users data. But the argument is not about that, it's about privacy, and how we have seen exceptions to privacy have always led to a sl
77.
▲
by
mythhabit
3y ago
That was a couple of months ago now. I left the site due to that, and just wondered if something else happened after.
78.
▲
by
mythhabit
3y ago
What happened "a couple of weeks ago"?
79.
▲
by
mythhabit
3y ago
Of course it's not regurgitation. But it is combining related concepts and I'm sure that somewhere in its training input, there is plenty of talk about increasing, decreasing, and constant entropy. However, it does not really say
80.
▲
by
mythhabit
3y ago
Talking about mixing coffee with sugar/milk and reversing the arrow of time (ie. entropy) is a widely talked about example. That said, I agree with your arguments and conclusions. Humans are not able to do this kind of reasoning until
81.
▲
by
mythhabit
3y ago
You didn't respond to this > Let's assume that in 10 years, LLMs are at the point where you cannot distinguish between a well written book by an author and one generated by entirely by an AI. Suppose we have two authors, one th
82.
▲
by
mythhabit
3y ago
I do not disagree with any of what you wrote. That is also an entirely different line of reasoning than your first argument. That said, LLMs today, cannot do this in a meaningful way. If an author cannot write a better book than ChatGPT, th
83.
▲
by
mythhabit
3y ago
I do not buy the premisses. Your example states that it can provide the full text of any book. An LLM cannot do that. They can produce something in the same style and setting. When an actual human author mimics other writers styles, it is n
84.
▲
by
mythhabit
3y ago
A car has the potential to be used for illegal activities. In fact happens every second, but that does not mean we go out of our ways to make them illegal. That said, it is still an open question if the content generated by LLM is illegal.
85.
▲
by
mythhabit
3y ago
You can, theoretically, keep constructing machines for every input you need. Yes you are bounded, but you can say "this machine will halt for every input it can receive in the next 100 years", because you need to quantify the inpu
86.
▲
by
mythhabit
3y ago
IO is just state and can be represented on the tape. You can construct a machine for every conceivable input at every time during the execution and reason about them.
87.
▲
by
mythhabit
3y ago
The naive way of deciding it it halts is executing it and keep track of all state between steps. If the machine halts, you're done, if it repeats a state, it's looping. If you can represent a program with an LBA, it is by definiti
88.
▲
by
mythhabit
3y ago
There is some nuances here. While the general halting problem for a general Turing machine is undecidable, and with a fairly easy to understand proof as well, the computers we run today are not a general Turing machine. They are of a weaker
89.
▲
by
mythhabit
3y ago
I actually think there might be an application for the current level of LLM to assist with requirements, changes, reprioritization ect. Imagine that all requirements, requests for change or reprioritization is routed though a LLM. Suddenly
90.
▲
by
mythhabit
3y ago
SQL is a pretty decent language. Sure it could be improved and made more intuitive and maybe even simpler. Relational data maps directly to objects if you have halfway decent tooling. SQL is almost 50 years old now - it'll be here for
More ›