3,769 karma · joined October 10, 2022
SFP transceivers are just that- transceivers, they will spit out what you put in, so you could even build some interesting bespoke fiber optic signaling gear with these, and a PCIe backplane, or just solo.
The margin on actual load to a permanent set is likely 2-3x what those trucks are carrying. And the margin to actual failure from just this quasi static loading and no external environmental loads (wind)... even more so.
Unless there was some catastrophic fuck up in construction process or calculations, this is completely safe.
I am definitely more aligned with needing what I would rather call 'Deep Semantic Search and Generation' - the ability to query text chunk embeddings of... a 100k PDF's, using the semantics to search for the closeness of the 'ideas', those fed into the context of the LLM, and then the LLM generate a response to the prompt citing the source PDF(s) the closest matched vectors came from...
That is the killer app of a 'deep research' assistant IMO and you don't get that via just grepping words and feeding related files into the context window.
The downside is, how to generate embeddings of massive amounts of mixed-media files and store in a database quickly and cheaply compared to just grepping a few terms from said files? A CPU grep of text in files in RAM is like five orders of magnitude faster than an embedding model on the GPU generating semantic embeddings of the chunked file and then storing those for later.
The average- nay, even the more above average journalist will never go far enough to discern how what we are seeing actually works at the level needed to accurately report on it. It has been this was with the technology of humans for some time now - since roughly the era of an Intel 386, we surpassed the ability for any human being to accurately understand and report on the state-of-the-art of an entire field in a single human lifetime, let alone the implications of such things in a short span.
LLM's? No fucking way. We're well beyond ever explaining anything to anyone en masse ever again. From here on out it's going to be 'make up things, however you want them to sound, and you'll find you can get a majority of people believe you'.
Fitting a -single- such surface to a set of points is nearly trivial; finding a way to best fit -multiple- such surfaces together to approximate a non-trivial shape (cloud of points) where they share edges in a way that could be joined like this paper model.... feels very NP-hard to me. This is a subset of the problem in the 3d-scan-to-CAD industry where you have a point cloud/mesh and you need to detect flat planes, cylinders, fillets, etc of a 3d scan and best-fit primitive surfaces to those areas and then join them into a manifold while respecting a bunch of other geometric and tolerance constraints.
There is a reason why there are only a few software packages that even attempt to do this, and it is almost always human-guided in some way. It's a fascinating problem.
It's happened to me several times where I doubt it could be simply recency bias coupled with chance of topics, due to the specificity and narrow directed-ness of the graph.
A random guy on another channel heard about how I was struggling to run certain things because my family couldn't afford anything newer than an old 486 with 8MB of RAM, on a motherboard that didn't support any modern graphics cards and at that point was barely able to run Windows. This person, whom I didn't know other than just a screen name and chatting about Paintball, just straight up sent me an entire motherboard, RAM, and CPU combo in the mail; A relatively new at the time AMD Athlon, on a motherboard with an AGP slot, so I could finally get a modern 3D graphics accelerator. Nothing asked in return, just 'here ya go kid, make good use of it'.
These two events when put together essentially changed the course of my life more than almost anything else, and both happened on IRC.
The game I was picturing in my head seemed far more exciting.
>Section 5.5.2: The “Spiritual Bliss” Attractor State
> The consistent gravitation toward consciousness exploration, existential questioning, and spiritual/mystical themes in extended interactions was a remarkably strong and unexpected attractor state for Claude Opus 4 that emerged without intentional training for such behaviors.
As in... what, AMD K6 / early Pentium 4 days was the last time I remember hearing about cpu cooler failing and frying a cpu?
Fun tidbit: The formation mentioned, the Five Towers, is right across the valley from where most of the movie "Cliffhanger" was filmed (the Tofane Group around Cortina d'Ampezzo).
The problem is it's not all text, you need the images, the plots, etc, and smartly, interstitially compressing the old stuff is still a very difficult problem even in this age of AI.
I have an archive of about 8TB of mechanical and aerospace papers dating back to the 1930s, and the biggest of them are usually scanned in documents, especially stuff from the 1960s and 70s, that have lots of charts and tables that take up a considerable amount of space, even in black and white only, due to how badly old scans compress (noise on paper prints, scanned in, just doesn't compress). Also many of those journals have the text compressed well, but they have a single, color, HUGE cover image as the first page of the PDF, that turns the PDF from 2MB into 20MB. Things like that could, maybe, be omitted to save space...
But as time goes on I start to become more against space-saving via truncation of those kind of scanned documents. My reasoning is that storage is getting cheaper and cheaper, and at some point the cost to store and retrieve those 80-90MB PDF's that are essentially total page by page image scans is going to be completely negligible. And I think you lose something be taking those papers and taking the covers out, or OCR'ing the typed pages and re-typesetting them to unicode (de-rasterize the scan), even when done perfectly (and when not done perfectly, you get horrible mistakes in things like equations, especially). I think we need to preserve everything to a quality level that is nearly as high as can be.
I keep about 16TB of personal storage space in a home server (spread over 4 spinning disks). The idea of expanding to ~200 TB however seems... intimidating. You're looking at ~qty 12 16TB disks (not counting any for redundancy). Going the refurbished enterprise SATA drive route that is still going to run you about $180/drive = $2200 in drives.
I'm not quite there as far as disposable income to throw, but, I know many people out there who are; doubling that cost for redundancy and throw in a bit for the server hardware - $5k, to keep a current cache of all our written scientific knowledge - seems reasonable.
The interesting thing is these storage sizes aren't really growing. Scihub stopped updating the papers in 2022? At honestly with the advent of slop publications since then, the importance of what is in that 170TB is likely to remain the most important portion of the contrib for a long time.
Either way, we don't know, and the original commenter doesn't know either, because they didn't google anything about e-bike battery sizes or their local electricity prices and thus didn't take the opportunity to actually learn something.
>GPT estimates (...)
(0.5 kwhr) * (20 cents/kwhr) = 10 cents. With an additional 10-15% due to charging system inefficiencies (lost to heat). 11 cents.
It can be good exercise to do an 8th grade level word problem every now and then.
The purpose being everything can be built from the bottom up and at no point is any part of it out of the scope of inspection. You never actually know what is spun into a that PIC64GX1000 because you didn't place and route and oversee the fab yourself.
Just because an ISA is 'open' like risc-v does not mean you know how it was implemented on-die.
Having all the verilog and using Yosys+nextpnr to generate the layout for the FPGA is (about) as close to knowing the provenance of the whole stack as you can ever get.
This line kinda got me down, because, well, last night I went out for a few pints and paid €16 for two drinks; Here we have a miracle of modern technology available shipped to your door for about the same price of what it now costs to just go out and do the thing people have done when socializing for the last 1500 years.
We're subsidizing the costs of all this modern tech by heavily taxing ourselves on the things once taken as nearly the bare minimum lifestyle.
e.g.:
It's been a long time since I've seen that (widely panned as not-very-good) movie, but I feel I remember a line about the little rover using an 'off the shelf computer modem' - this is actually true, the little rover communicated back to the Pathfinder base station with a straight up off-the-shelf RS232 9600bps wireless transparent modem link. [0] [1] I remember that detail as it showed that, even though the movie itself was... uhhh... interesting, science-wise, it clearly had someone in an advisory role that knew something about real JPL hardware on Mars at the time.
[0] https://urgentcomm.com/public-safety/data-communications-fro...
[1] http://www.iki.rssi.ru/mpfmirror/rovercom/itworks.html#rover...
Try to book a train trip from Brest, Bretagne, France to Vienna, Austria. Just halfway across western Europe. You'll wind up having to take four trains and likely change stations in a way that involves a metro or a bus, book on two or three different websites, if it goes perfectly it will take anywhere from 15-18 hours of travel time and if any single part gets delayed you will be staying in a very expensive overnight hotel and be out hundreds of extra euros.
Or you can take an Air France booking for literally half the cost, with a single plane change that is guaranteed, and it will take about 6 hours including the airport security and loading time.
Seriously, people constantly talk about train travel (in europe) like it is amazing and it's only true for very short distances. Otherwise it is a travel nightmare. And I hate airports too, but, things have gotten worse train wise in the last 20 years, I swear.
Finding that piece of software around 2001-2002 was what allowed me to finally download a specific piece of, ahem, 'shareware', that was about 400 MB, zipped, that I would never have been able to finish on a 14.4kbps modem on a single very noisy phone line that usually dropped the call every 2 hours or so. It eventually took three days but the file came across uncorrupted. It wouldn't have been possible without the ability to resume downloads after dropped connections.
And that software download went on to allow me to start the path learning what I wanted to learn about, and that paved the way for my engineering degrees and thus setting me up for the last 20-some years. Wild how little pieces of the puzzle like that drive so much of your life.
and I was surprised that no one brought up the very real downsides of being left handed in a left-to-right writing system region of the world (which is most of it). Most comments were leaning towards backwards conservatism and straight up malice with regard to students being forced a hand in writing early in school and it seemed no one brought up the very real practical reasons for preferring to write right handed, especially with ink.
And I say this as someone who is completely ambidextrous when writing but does not do the 'hook hand' left hand to write, and thus I usually write right-handed with pens and pencils. I have a left handed friend who does write that way and it just screams RSI/Carpal Tunnel to me.
Due to the inherent fuzziness/diversity in all models right now I don't think there is a universal approach to this idea but it is something people deploying these systems may want to try and detect.
It's like four pages and it lays out the fiction better than many 400-page novels or entire films dealing with similar subject matters have, IMO.
A melt of 1,000,000 pennies would be about 3000kg, so 510 kWhr of energy, and using an electric furnace that is about 50% efficient at putting its heat into the melt, that is about 1000 kWhr at the wall outlet. I've been in foundries with crucibles that could do probably about 200kg copper pours, with a furnace power of about 100kW, you'd do a full crucible melt every few hours, and you'd run this continuously until you're out.
If you live somewhere with electricity, say, $0.15/kWhr energy, that is a melt cost of $150.
1,000,000 pennies at 3 cents per worth of copper would be about $30,000 in copper reclaimed over the course of a few days.
So, it would be absolutely profitable in large quantities, assuming you ... already work in a foundry.
So, modern pennies are just copped plated zinc. But up to 1982 (barring some small interruptions like the 1943 steel penny) they were 95% copper. There was roughly 2.95 grams of copper in each penny.
Current copper prices seems to be approximately $12.31/kg or $0.0123/gram. That is about $0.036 per pre-1982 penny. Round down to 3 cents/penny to account for recycling costs. If you took all pre-1982 copper pennies you could out of circulation yourself you could make a 3x return!
(I have no idea how many pre-1982 pennies are still in circulation and it seems like this isn't public knowledge. Something tells me this endeavour to recycle them, as an individual, would never be fruitful. Also... my memory is that there is some question of legality in destroying currency en masse...)