Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
xyzzy123
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
42 ms
·
721.
▲
by
xyzzy123
3y ago
Note that asymptotically, believing that work is worthless is the same as ignoring it. I think this is interesting because in theory there is an infinite amount of potential work that everyone is ignoring at any given point.
722.
▲
by
xyzzy123
3y ago
Death panels are hard work and if you haven't noticed you're on one then you're shirking. (Less snarkily, we're all making resource allocation decisions constantly and on the margin those are life or death decisions). IM
723.
▲
by
xyzzy123
3y ago
mTLS tends to add infrastructure complexity, I much prefer it when you can "terminate" auth where you like in the ingress path for user-facing stuff rather than first-hop auth. I wish there were "real" standards and a be
724.
▲
by
xyzzy123
3y ago
If someone can show GPT-4 is "reasoning" (for some meaningful definition of that) in specific scenarios, surely counter-examples do not disprove this.
725.
▲
by
xyzzy123
3y ago
Gag orders are rarely used for this purpose. The information that so-and-so parties provided some information (without disclosure of that information) in response to a lawful request will usually not predudice a trial. What gag orders are f
726.
▲
by
xyzzy123
3y ago
IMHO facts "come to life" with more context, e.g: * How do we know that fact / how was it measured or determined? How sure are we that the fact is true? How accurately / precisely do we know it? * Who figured it out 
727.
▲
by
xyzzy123
3y ago
My kids go to a primary school which operates a bit like this, in Melbourne. It's about 30 kids / 20 families ish (it's hard to practically scale consensus decision-making beyond certain limits). It's operated as a paren
728.
▲
by
xyzzy123
3y ago
While current LLMs aren't really strong enough, I wouldn't completely write off the "LLM goes brrrr" approach vs domain-specific optimisations. How much effort is it worth investing in domain-specific tooling / lang
729.
▲
by
xyzzy123
3y ago
Now I'm really wondering if that's true or not. Software engineering can be self-studied very cheaply, lots of free resources, mostly you just need time and motivation. Failure is usually cheap. Farming on the other hand requires
730.
▲
by
xyzzy123
3y ago
If you model only the first-order effects (supply and demand of labour as a good, only) then sure, price goes up, it seems pretty clear that demand will go down. If you model the economy in more detail (with stocks and flows of money, and w
731.
▲
by
xyzzy123
3y ago
How do we even define a robot for tax purposes? Is a CNC machine a robot? A 3d printer? An injection moulder with pid control, a cnc box folder, a printer? a coffee machine? All of these can be considered "computer controlled machines
732.
▲
by
xyzzy123
3y ago
This seems disingenuous, afaik the contributions of GoF work (that enhances transmissibility, pathogenicity or immune escape of a virus) to widely deployed vaccines is minimal or zero. EDIT: if downvoter(s) could provide specific examples o
733.
▲
by
xyzzy123
3y ago
With similar excess deaths in Australia, I feel like attributing these to power prices is a bit of a stretch.
734.
▲
by
xyzzy123
3y ago
In order to be usable, many (most?) providers will implement some kind of override to allow recovery workflows. Have a look at https://media.fidoalliance.org/wp-content/uploads/2022/03/Ho... and multi-
735.
▲
by
xyzzy123
3y ago
To save anyone else the lookup, djb was not quite 19 yet :)
736.
▲
by
xyzzy123
3y ago
Also, diversity means you trust all of the providers; from a security perspective it's worse.
737.
▲
by
xyzzy123
4y ago
I think we might see specialisation, but driven more by proprietary datasets, fine tuning or specific supporting software. That is, a sort of protectionism and/or moat application, rather than limits on how many fields the LLM can be p
738.
▲
by
xyzzy123
4y ago
Have you tried recently? I am getting surprisingly good results lately with GPT-4 on basic physics problems with no prompt engineering (just stating the problem). It seems to recognize when it needs to "think step by step" and jus
739.
▲
by
xyzzy123
4y ago
Interestingly if you give GPT-4 the problem but append the instruction "answer only, no explanation" it gets it wrong (I got "108" as the answer in that case). It is fascinating how much it needs to work step by step.
740.
▲
by
xyzzy123
4y ago
Fundamentally I do think "operational rigour" is separable from "being a nice place for humans to work". But I don't think any discussion on this kind of thing is complete without understanding the power relationshi
741.
▲
by
xyzzy123
4y ago
It's a signal that correlates with risk of default on the loan.
742.
▲
by
xyzzy123
4y ago
Feels like you could make a case either way, like how "freedom fighters" or "resistance fighters" might be used for some groups vs "illegal combatants" for others... or even the same group at different times. T
743.
▲
by
xyzzy123
4y ago
How often do most people physically see CEOs, though? For most purposes maybe it's enough to project charisma in a video. An AI CEO might also be capable of superhuman "interpersonal" feats like "personally" answeri
744.
▲
by
xyzzy123
4y ago
Microsoft have a pretty considerable advantage - a valuable production workload to optimise and strong knowledge of what that looks like. Sometimes it's better to be the fast follower.
745.
▲
by
xyzzy123
4y ago
I think deep down, everyone already knows what the "killer application" for open source AI video generation is going to be.
746.
▲
by
xyzzy123
4y ago
Hi, what are the differences in this case between "abandoned" and "finished"? Guessing a lot of toolchain / software rot?
747.
▲
by
xyzzy123
4y ago
Mozilla or Wikimedia foundation have budgets with the right number of zeroes to support an effort like this. While those organizations might not be the right fit, I think they serve as an "existence proof" that a large scale open
748.
▲
by
xyzzy123
4y ago
Doctors have an unusually strong professional monopoly. In much of the world they regulate to maintain perpetual under-supply and to ensure they are inserted as bottlenecks in strategic/lucrative places. When AI comes for health care y
749.
▲
by
xyzzy123
4y ago
In anglo countries at least, population growth and a large portion of economic growth has been driven by migration. There are very strong incentives to try and "keep the party going", especially in economies (say Canada/Austr
750.
▲
by
xyzzy123
4y ago
I use it exclusively thru the terraform provider which abstracts over some of the details. I like it a lot more than cloud-init in this scenario.
More ›