The internet is for everyone. I am proud for us for building something so awesome that we get to train an entire replica of human reasoning on top of it. It's sad that most of the "new" internet will be made by these machines, but that's cool nonetheless.
Yes it's billion dollar companies building it, but every technical revolution needs large funding before it becomes accessible. Even the internet itself was way too expensive back in the days. Now we access it from fridges and toasters. Electric cars had to start as luxury purchase, so did phones or even CD players.
Now that we know what quantization is most optimal so that we built optimized accelerstors, how to architecture/harness LLMs for our purpose, now we can start to reclaim it.
Especially now when LLM APIs are starting to get expensive.
You still need some way to make the kernel to send those signals to the processes of your choosing. If the kernel decides to send SIGLOWMEM to xlock instead of firefox, the xlock will get killed because it really doesn't have any memory it can give up.
SIGLOWMEM, if it existed, would have no reason to default to anything other than IGN.
It would also logically be sent to “every” process on the machine, with the subsystem probably having a heuristic to skip processes which were already signaled and have not had significant memory increases since. The goal of an early warning is to cooperatively release memory (and maybe abort memory intensive computations), the kernel already kills processes “at random”, it does not need a second way to do that.
Looking at how GPU development has been sidetracked for NPU, I worry that this is 2035 target at best. Manufacturers will push for maximising matrix operation silicon area. In the era of trillion dollar investments into datacenters, traffic cost is afterthought. The only benefitors might be YouTube, Netflix and such, but on their scale investment into ISP level caches might be cheaper.
I used to love react of old with redux/flux and little to no hooks and weird side effects. It had super sound software design at the cost of being bulky.
I think we should return back to that with the era of LLMs. We don't write the code anyway. Let's use reducers and fully event sourced states.
What is the point of this comment? Do you mean to say we should cease the pursuit of DEI because we've "come so far"? Of course that's engaging with you with the assumption that you're not intentionally ignoring the weaponization of the movement in the US to further the goals of the current administration.
That would be a strawman. I am not living in US and don't care about US politics other than they are rising my diesel prices.
I'm just pointing out that historically this is the most DEI friendly we have ever been. If americans would be able to see outside a 4 year election cycle and realise there is a whole world outside the TDS zone, we could have a more civilised conversation than throwing strawmans around.
Because I feel like it's unappreciative to act like we didn't make any progress. It's really degrading to people that really did put in an effort (like myself). It's like comming to a celebration of a sport victory with "you could have done better" attitude.
So let's now focus on why you assumed a hidden meaning somehow connected to "current administration"? Why this knee jerk reaction?
We used it for payment processing. We got huge CSVs from various APIs and used string decimals for computing to avoid overflows/underflows and rounding errors.
reply