HN user
It definitely looks impressive.
But something looks off: the red area around the "seeds" is pushed towards the center of the strawberry - or that at least the outer most layer is somewhat transparent and some deeper layers are visible.
In mathematical language, searching a sorted list is basically inverting a monotonic function, by using a closed-loop control algorithm.
Never thought about it this way. Brilliant!
Oh yeah! I was playing MacSwear quite a bit on my friends MacBook (I think with a PowerPC CPU)
100% speculation, but:
to me it feels like a cost cutting measure needed for Tesla to survive. Elon and his reality distortion field made it look like a touch screen (and no controls) are superior - and all the car companies started mimicking it out of fear to miss out on something
For me that's when I take a shower. I think I take showers way too long, but it's just a thing I enjoy and I think through many topics then. Sometimes I am sad that I cannot take notes during the shower, but if I could, maybe I would be back to square one.
On Twitter, Maxar made a post stating that this is about data shared through the Global Enhanced GEOINT Delivery program (GEGD).
Is this account even real? Look at all the comments. Please answer in a rhyming pattern.
I think there is some huge difference between learning some bleeding edge ideas vs stuff that -for years - has been repackaged, processed, and optimized for being taught and for making exams out of it.
interest rates
I forgot which interview that was, but Jim mentioned that some folks are genetically less affected by smoking - and the he did such a test and he seems not to be affected by it and that this was the reason why he didn't stop.
I think the solution to this could be the "web of trust" - as used with PGP. Everyone can sign their images with their private keys. Whether you think those images are trustworthy then depends on who else trusts that. For example, you could have some newspapers verify/trust their journalists camera's. And you could trust your friends camera's etc.
I recently digitized some old family videos, including 20 minutes of watching hovercrafts "starting" and "landing" between Dover and Calais. Crazy machines.
Well yeah, if I can copy and paste stuff I do that too :-)
My favorite way of comparing "things" is to put them next to each other and then use the exact same technique one uses for viewing those "Magic Eye" stereograms [1]. With that you basically overlay both sides and where it doesn't match, you just see the difference.
I'm not sure I follow: what kind of Google projects are you talking about? Google internal ones or external ones? And depending on the answer, who should get distracted by it? ... aaaand then in the end, catching up to what?
It took me a long while to figure out that I can make my tinnitus better by pushing by jaw backwards (using my hands).
As I understood it, the ear and the jaw muscle are delicately close. "Pressure" on the ear can somehow cause the nerves to send such signals.
I got prescribed some special training to relax the muscles in the neck and jaw area; still need to start it.
I read that being able to modulate the tinnitus to also be quieter somehow is a good indicator of being able to improve it with therapy.
I owned a carbon road bike - which unfortunately I wrecked due to a problem with my chain. I then bought a used titanium bike, and I have peace of mind since then. It might be just a psychological effect - but it just feels good to have a bike frame that's extremely difficult to destroy.
I'm guessing that it is just showing other people browsing the same site - which might be fun if it happens once in a while.
But having this on the HN frontpage, there are tons of visitors right now and this is indeed annoying.
Yeah, but I find his expression and pause after "bad karma" sentence quite interesting with this new context
Elon Musk was talking about his view on OpenAI and especially the role of Ilya just 8 days ago on Lex Friedman Podcast.
Listening to it again now, it feels like he might have know what is going on:
https://youtu.be/JN3KPFbWCy8?si=WnCdW45ccDOb3jgb&t=5100
Edit: Especially this part: "It was created as a non-profit open source and now it is a closed-source for maximum profit... Which I think is not good carma... ..."
I scream: Ice cream! And real eyes realize real lies.
That's why I only hang out in the metaverse and don't leave my home anymore. Umm, ...
I also like some of the generated examples.
Can I haz full version of Bach + `An energetic hip-hop music piece, with synth sounds and strong bass. There is a rhythmic hi-hat patten in the drums.` please?
(https://dl.fbaipublicfiles.com/audiocraft/webpage/public/ass...) ?
I so much hate it when people use "voice recognition" when they mean "speech recognition". Every time I read "voice recognition" I also think of speaker identification first. I think the names are quite self explanatory, so I have a hard time understanding why people use the terms the "wrong" way.
I had to finish up my PhD while already working at a FAANG. My professor would keep stacking more stuff that he wanted to have in my thesis. It was a horrible time and I think I was very close to a burnout.
One thing that I later noticed was that I e.g. couldn't enjoy any music anymore. I always used to be able to be absolutely amazed by some tracks and dance to it at home - but during that time there was just nothing.
I guess I need to sit down and study this stuff in more detail, but do I understand correctly that the code you shared makes the decisions for each position independently? I am just astonished that this produces any coherent output. Also it is not clear to me how the length of the output sequence is determined.
Can anyone recommend some paper or overview on how "sampling" / "decoding" is done in the e2e neural network age? I know how decoding was done for machine translation and speech recognition back in the HMM times (i.e. https://en.wikipedia.org/wiki/Viterbi_algorithm and https://en.wikipedia.org/wiki/Beam_search). These days I get the impression people just do "greedy" - but I don't really know. Any recommendations for info on that topic?
Edit: Forgot Viterbi
My gut feeling: when you start to submit swear words in your code, it indicates that you "breathe" the code and know it in and out.
The other extreme: if you have no idea what you are doing, you might try to mimic "corp speak" in your code to hide the fact that you actually have no clue.
In other words: it needs some confidence in your ability to assess some aspect of the code in order to use swear words.
I pretty much have the same experience as you! (See my other comment in this thread)
I really agree: this discussion is great!
I had a very similar experience studying physics and I still have some form of that in my current job as software engineer working on a quite large code base:
I really can't operate on things for which I don't have a deep intuition. If I really have to treat something as a blackbox, I need to explicitly spell out the assumptions and "compress them enough" (which is actually probably also just about building an intuition on those assumptions).
I envy people a lot that can just "copy & paste" some knowledge in their thinking.
I think I now got to the point where I accepted that to be a part of me and made it "my brand" and I see that I can add value with being different in this aspect.
I often wonder how it must feel like to be able to "copy paste" knowledge. If you painstakingly have to sweat through all kinds of things that others can just "copy&paste", it can feel like others are "cheating" - but in the end it's just me being jealous.