Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
RhysU
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
33 ms
·
451.
▲
by
RhysU
3y ago
Thanks for noticing the implications. Your great-grandkids would find utility, as would their children's children. It's their perspective that matters for utility, not yours. The fun thing is people talk log utility but forget the
452.
▲
by
RhysU
3y ago
When moving, they take up the space of about 2 bicycles riding safely. When parked, about 4 bicycles. It's a congestion tax not a fumes tax.
453.
▲
by
RhysU
3y ago
Your heirs/descendants wouldn't be.
454.
▲
by
RhysU
3y ago
Motorcycles don't cause congestion. Motorcycles can travel through congestion when permitted to lanesplit.
455.
▲
by
RhysU
3y ago
This product release marks the beginning of the Spacer/Settler divide.
456.
▲
by
RhysU
3y ago
Hiding a backhanded dig ("anything of substance"?!) is worse. Just call bullshit in return. I wear big britches.
457.
▲
by
RhysU
3y ago
A definition of "plausible" is "apparently reasonable and credible, and therefore convincing". In what limit does "apparently reasonable and credible" diverge from "true"? We'd make the LLM not
458.
▲
by
RhysU
3y ago
> For a LLM, it is to generate plausible language. LLMs are not being sold as delivering only plausible language. Is it the craftsman's fault when the salesman lies?
459.
▲
by
RhysU
3y ago
I conceptually agree with you that a fool blames his tools. However! If LLMs produced only lies no one would use them! Clearly truthiness is a desired property of an LLM the way sufficient hardness is of a bolt. Therefore, I maintain that a
460.
▲
by
RhysU
3y ago
> Large Language Models (LLMs) are never wrong, and they do not make mistakes. I call B.S. If LLMs never made mistakes we wouldn't train them. Any random initialization would work.
461.
▲
by
RhysU
3y ago
(Sure, no compile time. But how much time do you think humanity collectively wastes waiting for CPython to import things? On. Every. Run.) "Did it work?" is a point-in-time observation. It is always situational. To find out, I hav
462.
▲
by
RhysU
3y ago
My favorite version of this observation is asking tenured professors how they picked up their sweet digs. Often the answer is "I knew I would be here indefinitely. I saved cash until the bottom fell out. I bought."
463.
▲
by
RhysU
3y ago
> The whole point of Python is the duck typing. Duck typing implies that "Should this work?" has one true answer: "Did it work?" I am at work to work, not to play guess-and-check with upstream libraries.
464.
▲
by
RhysU
3y ago
There's a comical Jurassic Park "life finds a way" outcome lurking... those reproductively restricted by these ideas gradually will be replaced by those who are not so inhibited.
465.
▲
by
RhysU
3y ago
How did we go from discussing the step up basis to assuming I want an "unhealthy modern society"?
466.
▲
by
RhysU
3y ago
I agree. But if we had a simple system people would be able to tell whether or not they get adequate value for their money. Obtuse and opaque are a defense against fair or just, take your pick.
467.
▲
by
RhysU
3y ago
> It’s a heads I win tails you lose approach. All capital gains taxes are heads-govt-wins. The step up is simply not applying that insanity, for once, at death. That is, it's marginally sane.
468.
▲
by
RhysU
3y ago
The step up applies only to assets in one's estate at death, not assets in trust. IIUC, the very wealthy have limited exposure to step up in basis because they use trusts. Everyone else doesn't want to owe taxes when selling a hou
469.
▲
by
RhysU
3y ago
I love that such systems were built once: https://www.wikipedia.org/wiki/Setun
470.
▲
by
RhysU
3y ago
Gold is valuable when mined at 0 grams/second. Is Bitcoin valuable at 0 transactions/second? If not, how much slower than 7 transactions/second is acceptable? Why?
471.
▲
by
RhysU
4y ago
> If I were a 20-something today I’d stay away from high cost cities. Disagree. Saving X% of one's early career salary goes a lot further when the salary goes up near a city.
472.
▲
by
RhysU
4y ago
Definitely. I just remember the pages of total junk article results starting about 2010.
473.
▲
by
RhysU
4y ago
Would the old way have worked circa 2010 before SEO articles ruined vanilla search? Is the new way preferable because SEO-ish approaches haven't caught up? If so, can we limit searches to the LLM's trusted learning corpus and get
474.
▲
by
RhysU
4y ago
The USB cable to the phone pesters me to install/enable Android Auto every time.
475.
▲
by
RhysU
4y ago
Occasionally, a CD player.
476.
▲
by
RhysU
4y ago
I don't want CarPlay or Android Auto. It's one more thing I have to fiddle with. One more thing to fight while someone beside me or behind me becomes impatient. Bluetooth is meh. I would rather have an aux cord. From the days when
477.
▲
by
RhysU
4y ago
I vote for 2-stroke. We certainly burn oil.
478.
▲
by
RhysU
4y ago
This. Everyone should pay quarterlies and employers should not withhold. People should feel the visceral pain of having their money taken away and be inconvenienced by the complexity of the tax code. Then we can have real discussions abou
479.
▲
by
RhysU
4y ago
No, the implication is that most of us fake it until we make it. And The Peter Principle says we're all always faking something. My comment was just about humanity. ChatGPT isn't worth writing about.
480.
▲
by
RhysU
4y ago
Most of us are much worse on coding problems not in our training set! (Looks down at dynamic programming problem involving stdin/stdout and combining two data structures).
More ›