I said favorable circumstance, not favorable conditions. I mean the person in front of you doesn't stop quickly or veer into your lane very often regardless of whether there is e.g. fog or some confusing signage. Knowing what to do in a confusing situation is a hard problem, but avoiding a crash when someone stops abruptly with full visibility requires mostly safe management of speed, distance, and lane exit options, which people are terrible at.
The amount of non-determinism in compiler and LLM output is so vast that even comparing them is silly.
Tests are not a one time thing. They need to be maintained, and if you use same LLM to do it, it will happily delete failing ones, so better use a seaparate agent, and be prepared for the bill.
> The amount of non-determinism in compiler and LLM output is so vast that even comparing them is silly.
What is that supposed to mean? Yes, LLMs are often considered to be non-deterministic because, in practice, they rely on systems with unstable floating point rounding. Likewise, compilers are often considered non-deterministic because, in practice, they often rely on systems with unstable thread execution order. Of course, those are both just implementation details. You can, albeit at the cost of things like performance, remove the hidden inputs. Computers are fundamentally deterministic! Anything that runs on computers must also be deterministic.
If you were to run a compiler or LLM on a Turing machine, they would both be 100% perfectly deterministic. We only see them as being non-deterministic because, in the real world, we run them on hardware that plays tricks with computation by introducing inputs via hidden side channels. But, again, those are implementation details. Compilers and LLMs are not defined by any particular implementation.
LLM inference can, with a bit of know-how, made 100% deterministic.
Of course tests need to be maintained. They have always been. And the test plan you write determines what and how it needs to be tested, just like it has always been.
When I listen to some people or to stuff that comes out of the frontier labs, they indeed do sometimes sound like a religious cult talking about deity. I find it disturbing.
Ventilating at night doesn't work at all when nights are hot. And now it happens oftenly, not sometimes. That is a bigger problem than hot days.
Also, all the activity you do inside your insulated home also increases temperature. Human body itself is a small space heater. At some point, you gotta get that energy out somehow, and if nights are hot, there is no way to do that except via a heat pump.
When it is 40C max by day and 22C min by night, for example, the outside is still cooler than the inside at night. So ventilating will have a real impact.
With this you might reach 25-26C inside in the morning and 29C by the end of the afternoon. It is hot but generally OK with fans, depending on humidity.
Then you can turn the AC on during daytime on top of that, which makes more sense than relying only on it.
I mean, fortunately for me it's dry enough that I can run a small swamp cooler for the couple of weeks a year where those are the temps in my place. But yeah, you're correct that this is what it looks like where I am at, temperature-wise.
Though I have no A/C, per se... I have thought about it. I'm 100% off grid, but the only time I would like that is when the sun is so far up that the batteries have all already charged.
reply