Alignment: neutral evil
Characterization: possibly obstinate, certainly cantankerous
Citation: https://en.wikipedia.org/wiki/J%C3%B6rmungandr
- >Deepfakes have immense creative potential
...and the lawyers win. =3
- >Can we agree that's, in general, a good thing?
The models come from overt piracy, and are often used to make fake news, slander people, or other illegal content. Sure it can be funny, but the poison fruit from a poison tree is always going to be overt piracy.
I agree research is exempt from copyright, but people cashing in on unpaid artists works for commercial purposes is a copyright violation predating the DMCA/RIAA.
We must admit these models require piracy, and can never be seen as ethical. =3
'"Generative AI" is not what you think it is'
- This covers the data center resource green-washing rhetoric, and most taxpayers will be paying more for energy now regardless of what they think:
'"Generative AI" is not what you think it is'
https://www.youtube.com/watch?v=ERiXDhLHxmo
And this paper proved the absurd outcome of the bubble is hype:
'Researchers Built a Tiny Economy. AIs Broke It Immediately'
https://www.youtube.com/watch?v=KUekLTqV1ME
It is true bubbles driven by the irrational can't be stopped, but one may profit from peoples delusions... and likely get discount GPUs when the economic fiction inevitably implodes. Best of luck =3
- The studios did already rip off Mark Hamill of all people.
Arguing regulatory capture versus overt piracy is a ridiculous premise. The "AI" firms have so much liquid capital now... they could pay the fines indefinitely in districts that constrain damages, and already settled with larger copyright holders like it was just another nuisance fee. =3
- The encrypted page memory manager hardware in some ancient Sun systems prevented a lot of these context isolation problems. However, the modern IT landscape chose consumer grade processor architecture and bodged GPUs as the cloud infrastructure foundation.
Thus, there currently is economic inertia entrenching vulnerable system design. I don't think there is a company large enough to change the situation anytime soon, as the market has spoken. =3
Rule #3: popularity is not an indication of utility.
- Almost as bad as the Intel Management Engine. lol =3
- Could always try:
echo 1 > /proc/sys/kernel/modules_disabled
Which is supposed to block dynamic loading modules until a reboot. =3
- Most production OS I saw would do this on boot-up completion:
echo 1 > /proc/sys/kernel/modules_disabled
Which is supposed to block dynamic loading modules until a reboot.
It would be interesting if the PoC can get around that trick too. =3
- If a person is profitable to have around, than the conversation is simple:
1. Company has money people may want
2. People have unique skills a company needs to profit
3. Everything else is 100% unrelated BS
On average, every software developer brings in >$1.3m USD/year in additional revenue. If you are being exploited, than just find a better gig someplace better... as it is usually easier than advancing out of a critical role.
Salaries with legal encumbrances are usually just a terrible deal in the long-term. =3
- Anyone that actually works with CSS will tell you it is the platform specific standards mess that made it a problem. Try something fancy, and it will likely become a bug later.
It gets painfully clear if you want pixel-perfect layout on every desktop screen, mobile screen, and browser. YMMV, and PDFs will be around forever now... =3
- >I've always wondered what happened to him.
Probably, just ran out of Quarters for awhile... or discovered Final Fantasy VII on PlayStation. =3
- JLPCB White Jet Process (WJP) service starts at $5, and is full color including clear.
Home Resin Printers should be in a vented garage, as the unpleasant odor is the least of its issues. The low-viscosity washable water-like resins tend to be much easier to handle, but even when vented outside... an activated charcoal filter is wise if you have neighbors. =3
- Yet there was always that one kid that knew how to soft-lock Street Fighter II arcade cabinets with Guile. Samurai Shodown, The King of Fighters, and Mortal Kombat were also fun. =3
- From the Linux users choice of:
1. outdated legacy OS that has to rely on a limited subset of flatpack, snap, containerization, and CVEs...
2. modern OS with unfix-able driver issues, NVIDIA wanting people to buy a new GPU, and broken abandonware drivers
I'd say a bricked GPU is appropriate terminology, as your screen will be forced into 800x600 by the hardware recovery mode and be functionally useless. For Linux users the only solution is to buy a compatible GPU, and or throw away the laptop/tablet/phone devices that can't run legacy Windows drivers.
When the collapse cycle of spiral development happens: a lot of Applications are also lost in the process, as the original publishers may not be maintaining an application a decade later. dkms with mystery Blobs would fall into the same out-of-band classification.
It is funnier people are upset about slop articles nonsense assertions.
The question about whether something is qualitatively "good" is meaningless if its demonstrably broken. =3
- Another straw man fallacy is silly. And "all software is terrible, but some of it is useful."
NVIDIA abandoning legacy drivers had to happen eventually, but it isn't an excuse for Linux bricking the hardware. The vestigial deprecated areas of the kernel is covered in foundation courses to help explain what mess to avoid, as the rule to never break user space hid a lot of the uglier bodges. In terms of code volume, linux source has been mostly driver code for quite a few years.
Trying to mix commercial blob binaries in a FOSS project is rarely stable over the long term (hence why recent RTX and AMD open drivers is very important.) Yet how many android devices are in garbage or recycling heaps? Answer, _all_ of them eventually, and usually in less than 3 years on average.
Linux will also soon no longer include 32bit x86 support if upstream planning goes as planned. It also had to happen eventually.
It should be noted the CUDA enabled drivers on windows is quite a bit more complex than the windows kernel itself. Yes it is still an awful adware OS, but it will still let you run 10 year old titles mostly without an issue. YMMV =3
- Back-porting modern drivers/security patches is another issue, but most static-linked applications tended to work fine up until Windows 11. If you wanted to play legacy games in old Windows environments it usually works up to the final patch level of the product. Note active Windows Service Packs meant the release was still supported, and most companies would maintain driver compatibility if and only if they were still in business.
Abandonware popularity is a relatively recent phenomenon, as hardware became cheap enough to be disposable... and SaaS/DRM hit the entertainment industry.
Most people are not going to port AAA title dependencies to ancient hardware, but the "Doom will run" on just about anything meme is usually still true. =3
- Some communities tried to keep EOL legacy nvidia drivers viable on newer OS kernels, but it is a poor allocation of resources prone to failure.
It is not FUD, but rather the consequences of out-of-band proprietary mystery blobs, dependency injection, and major structural changes within the kernel or user space programs. Many legacy dkms simply don't survive the code permutations over the long term (usually Wifi cards, equipment, and GPUs.) Thus, projects relying on such drivers break eventually as people start to abandon the legacy platforms.
Linux is good at many things, but LTS only slows the compatibility decay cycle to years instead of weeks. The several thousand tonnes of waste hardware it turns into garbage, and locked offline Application licenses do matter to some folks. It is the hidden Spiral development liability in most FOSS projects. =3
- I have yet to see a fully working nouveau based system, as most people are just looking to blacklist the mod given the collateral problems.
Depends on the hardware codecs use case, as some legacy cards are valuable to people that own legacy workflows.
Without CUDA + hardware-encoders a GPU is just a paperweight regardless of age for some use-cases. =3
- note 580 is now also EOL, and a rack of hardware video encoders is less demanding than gaming. =3
- Slop content is still slop... especially when saying things people want to hear.
Fact is most modern Linux Kernels >6.8.x don't support the legacy nvidia driver 470. Thus, on modern OS distros a lot of legacy video cards and laptops running Linux >6.18.x just isn't practical without a GPU upgrade.
The situation will probably get worse in the next few years as perpetual permutation kernel culture hits unmaintainable EOL hardware driver Blobs.
Most dual boot a Windows 10/11 ssd, as it still sort of works. YMMV =3
- Unless one has a rack of older GPU hardware that uses an abandoned EOL NVIDIA kernel driver difficult to install past kernel 6.12.x Then one faces the harsh reality of Windows users rightfully laughing at a perpetually Beta Linux OS, as Win11 still boots with the older drivers. while the dkms build randomly implodes at some point.
People dual boot SSD OS for very good reasons, as kernel permutation is not necessarily progress in FOSS. Linux is usable by regular users these days, but "Good" is relative to your use case. YMMV =3
- Most know some version of the 4 types of customer in marketing, but only a few figure out 3/4 classes of sales are not worth the effort.
1. The miser: No matter the cost, the right retail price is $0. These folks make up 82% of the market, but are usually effectively irrelevant in terms of revenue. Yet if you sell low-end low-margin products, than these are your customers.
2. The technical: Give them a list of specifications, and leave them alone. These people already know what you have for sale, and probably know the product better than most of your team. Too bad, these folks are <3% of the market, and while they have opinions they also don't matter in terms of revenue.
3. The sadist: These people are only interested in making people miserable, and for whatever reason are always a liability to have around even in the rare event they buy something. At <5% of the market they are also irrelevant in terms of revenue, but will incur additional losses though nasty cons etc. Your best bet is to give them free swag bribes, and refer them to a competitor because they are so awesome.
4. The emotional: These people are the highest profit market, as they are more concerned with how they feel about a product or brand. They don't care much about hardware performance specifications, but rather focus on the use-case in a social context.
One may disagree, but study 23000 users buying habits... the same pattern emerges for just about every product or service. Note, these classes are only weakly correlated with income level.
Thus, depending on the business it is absolutely possible to ignore the vast majority of the market while still making the same or greater profit. Yet if a product is mostly BS, than the online communities will figure that out sooner or later. =3
- Empty platitudes from an LLM will now likely increase in frequency. =3
- Sounds like a GIS company may be a good fit, or another related service. =3
https://www.esri.com/en-us/about/careers/job-search
https://www.jpl.jobs/c/educational-jobs
https://www.spacex.com/careers/jobs?programs=Starlink&discip...
- The STL file literally means Stereolithography, and was created by 3D Systems. It is the intended use-case, and is limited like any other file format.
The fact many fabs avoid DXF is unit and format version ambiguity. However, the other reason is predicting bend deduction is often wishful thinking. Most high end CAD packages are fairly good at handling that mess for people, but often still have issues figuring out how features will deform.
CAD can also start to fail on large complex geometries, as even a simple single combo-tool g-code macro to drill, helical mill, and chamfer threaded-holes in a plate often is done by the time CAD/CAM path-planners finish/crash.
Best of luck =3
- These should be read at least once in your life if interested in building industrial grade electrical, mechanical, and or software.
1. https://nepp.nasa.gov/whisker/
2. https://standards.nasa.gov/standard/NASA/NASA-STD-87394
3. https://standards.nasa.gov/NASA-Technical-Standards
4. https://sma.nasa.gov/sma-disciplines/workmanship
5. https://www.stroustrup.com/JSF-AV-rules.pdf
6. https://en.wikipedia.org/wiki/The_Power_of_10:_Rules_for_Dev...
7. https://www.nist.gov/pml/owm/laboratory-metrology/metrology-...
8. https://www.mitutoyo.com/training-education/
9. "Memoirs of extraordinary popular delusions and the madness of crowds" (Charles Mackay, 1852, https://www.gutenberg.org/files/24518/24518-h/24518-h.htm )
The artifacts are usually beautiful from good Workmanship Standards, Design For Manufacturability, and systematic Metrology. Dragging us all into the future one project at a time.
Note that training an ML model with such data would be pointless, as statistical saliency forms a paradox with consumer product design compromises. Note, there are _always_ tradeoffs in every problem domain.
'What it actually means to be "AI Generated"' ( https://www.youtube.com/watch?v=ERiXDhLHxmo )
https://www.youtube.com/watch?v=iXbzktx1KfU
Have a nice day, and note >52% of the web is LLM slop now. YMMV =3
"Memoirs of extraordinary popular delusions and the madness of crowds" (Charles Mackay, 1852)
https://www.gutenberg.org/files/24518/24518-h/24518-h.htm
I look forwards to buying the failed data center assets. LLM make great search engines, but are not the path to "AGI". Neuromorphic computing looks more interesting. Have a great day =3