Joeri

22 ms·

241.
by Joeri 3y ago
600 mhz is quite the machine for NT. I remember running NT4 on a 233 mhz pentium II with 128 MB RAM and everything felt instant and limitless. Windows 2000 was quite the hog compared to NT4 and all it added that I had a use for was USB supp
242.
by Joeri 3y ago
TIL Anyway, the profile is accurate enough that the dimming of non-HDR parts while ramping up brightness is not noticeable to me if I block off the HDR part.
243.
by Joeri 3y ago
Actually the whole screen is HDR, but the non-HDR parts are dimmed to match their normal brightness. They can do this because they calibrate their displays and know exactly how bright they are. By the way, if brightness is not maxed this wo
244.
by Joeri 3y ago
As this is a government contract, and there are strict public transparency rules on government contracts, I went digging. Here's the call for tenders: https://etendering.ted.europa.eu/cft/cft-display.html?cftId=...
245.
by Joeri 3y ago
Huffman seems to be drawing inspiration from Elon Musk, according to a recent interview. “Long story short, my takeaway from Twitter and Elon at Twitter is reaffirming that we can build a really good business in this space at our scale,” H
246.
by Joeri 3y ago
I started with rectangle but now prefer moom. https://manytricks.com/moom/ It integrates in the way I would expect the native feature to work: hover the green button for a snapping palette.
247.
by Joeri 3y ago
This is known as the theories X and Y of management: https://en.m.wikipedia.org/wiki/Theory_X_and_Theory_Y The personality of the leadership will determine whether they think of workers as self-motivated (will actively
248.
by Joeri 3y ago
I just asked an LLM to create a language and provide a demonstration and this is what it said. Call it a stochastic parrot if you want, but I’m pretty sure a linguist can prompt it to properly invent a language. Sure, I can invent a new la
249.
by Joeri 3y ago
In a sense LLM's are a Rorschach test for people's beliefs about consciousness. If you believe consciousness to be an emergent phenomenon derived from simple deterministic biological processes, then it is not a big leap to believe
250.
by Joeri 3y ago
Apple has an app developer story for ipad with swift playgrounds. It would not surprise me that their long term strategy is to grow that into the default app dev experience and have it be on all platforms, including visionOS. As for develop
251.
by Joeri 3y ago
Bad people will procure guns or make them from parts or scratch. How many of the people who perpetrate school shootings had the skill to build their own gun, or the criminal connections to procure one illegally?
252.
by Joeri 3y ago
I’m not sure I agree, but I’ll do my best to argue the position. The neural net software algorithms have been around for decades. What made LLM’s feasible are the hardware advances to achieve unprecedented scale, just barely providing the a
253.
by Joeri 3y ago
It depends on what sort of community you follow. The ML and tech podcasting communities have largely moved over. The politics, journalism and celebrity part of twitter hasn’t moved over. The corollary to that is that much of the vitriol and
254.
by Joeri 3y ago
There may not be enough text content on the internet, but there’s plenty of audio and video content, and there has already been some research about connecting that as an input to an LLM. So far we’ve seen that the more diverse the training
255.
by Joeri 3y ago
The point of democracy is not that everybody is perfectly informed to come to the logically best outcome. The point is that it gives a people agency over their own future. If they choose to be ill-informed and pick bad candidates, they own
256.
by Joeri 3y ago
The prompt is everything. It’s sort of like knowing how to formulate a google search query, but cubed. You need to learn a whole bag of tricks and you can’t use the intuitions of talking to people for talking to the AI. Some people are by n
257.
by Joeri 3y ago
You don't need code really. All you need is the right prompt. GPT-4 should be well able to convert a description of the kind of personality and roleplay you're looking for into a suitable prompt for any LLM. As programmers we&#x27
258.
by Joeri 3y ago
What the skeptics seem to get hung up on is the lack of common sense. These LLM’s are trained on language describing the world, not on the world itself. They lack the lived experience that we take for granted. So it is trivial to ask them c
259.
by Joeri 3y ago
A mac mini running home assistant with some kind of llama plugin running one if the 7B models should do well enough and would run entirely locally. It is just a matter of time before someone builds it. I think one of the issues for a lot of
260.
by Joeri 3y ago
I would say: run with it. Assume the model's output is always untrusted. Assume that no rules can be applied to the model itself, only to the container it runs in. What kind of sandbox does the model need to be placed in given those as
261.
by Joeri 3y ago
With prompt injection all model output is untrusted. There is no way to avoid this. That means every action the model takes must be labeled whether it supports untrusted input or not. If it doesn’t then an explicit step to turn it trusted i
262.
by Joeri 3y ago
Actually I would say this is yet more proof of how awesome breath of the wild was, and the incredible strength of a good franchise. 10 million people instabought it just in the hope it will live up to its predecessor.
263.
by Joeri 3y ago
People will make sure it is embodied. Do any of the major powers have the luxury of turning off their superhuman AI when up against that of the other powers? Definitely not. People will willingly give up control, even as they see the AI gra
264.
by Joeri 3y ago
There are some threads with hardware recommendations in the LocalLLaMa subreddit. Here’s a recent one: https://www.reddit.com/r/LocalLLaMA/comments/13f5gwn/home_ll...
265.
by Joeri 3y ago
My experience is that even the "toy" 7B models can do pretty amazing things, but they're very sensitive to prompting in the right way to force the model down the right pathways. My assumption is that the larger models have mo
266.
by Joeri 3y ago
Should all addictive products be allowed? When does such a product deserve being regulated or outlawed?
267.
by Joeri 3y ago
Ligatures in a fixed width programming context still take up the full width of the separate characters, so a unicode symbol of the ligature has a different width from the ligature itself, and this is immediately obvious.
268.
by Joeri 3y ago
You can also make a global one for an entire shared folder. For example, I wildcard all node_modules folders as excluded.
269.
by Joeri 3y ago
Syncthing is excellent. I’ve used it for years to keep my development folders synced between mac, linux and windows, and except for some initial figuring out of ignore patterns and the occasional hiccup with filenames that windows can’t han
270.
by Joeri 3y ago
If that makes you uncomfortable you definitely should not go reading the evidence supporting the notion that conscious free will is an illusion. https://www.mpg.de/research/unconscious-decisions-in-the-bra...

More ›