1,753 karma · joined May 12, 2022
"A Generalist knows less and less about more and more until he knows absolutely nothing about everything."
Eventually the cycle/bubble collapsed, and the same 1930s tariff policy literally starved people for almost a decade.
I really hope we don't repeat that history. However, the Bear market distant roar can't be ignored. =3
Mostly seeking design tips for a vacuum tolerant micro-positioning stage project. =3
That was nodejs, and it turns out people unwilling and or unable to improve things through systematic skill refinement just degrades into chaos.
Successful ecosystems are just pedantically lame enough to keep silly folks from going YOLO, but empowering enough to allow people to still have fun.
>which model has the best training code that was Pac-Man.
You mean which model is more cautious about copyright bleed-through of the $9Tn in FOSS and user code they misappropriated though isomorphic plagiarism. =3
The hidden agent risk in LLM often can't be detected during training and evaluation. =3
You were lucky, as most university contracts clearly state they own all associated IP created while working on campus. They could have legally given you $0, and end of employment notice. These things can get messy fast. =3
A bit of cash can turn folks into awful people, and some contracts are just a obfuscated legal con of the naive. =3
>state of the art GCs
That just isn't true, most of the Subscription Enterprise Performance Pack features are under a paid license by Oracle. The FOSS options are always slowly catching up, or not standards compliant enough for some commercial programs.
Being a sentient turnip, I probably wouldn't understand such things. Yet in general, 3rd party dependency injection into a design pattern should be considered a long-term variable cost. =3
Rule #3: popularity is not an indication of utility.
https://blog.stenmans.org/theBeamBook/#_concurrency_parallel...
Rule #23: Don't compete to be at the bottom, as you just might actually win.
In production, Go has proven solid for several years. It is best when used with the native code people ported.
There are only two issues I encountered:
1. getting the legacy ancient C source meta-circular Go compiler working to port the Go boot-strap compiler upgrade chain is a kick in the pants. However, once it is on a architecture it has proven rather resilient.
2. memory limited systems can develop reliability issues, as Go programs will often ungracefully throw hard to diagnose unrelated errors during each crash. A good metric is 3:1 of your average load as a safety margin (if you see 2GiB in average RAM use, make sure to over-provision the host with 8GiB RAM etc.)
Other than the above short list of edge cases, if you join a pure Go project it is usually pretty reliable. Most community folks interested in the language seem fairly competent at building stuff that is fun. =3
My comment was advice that most people will learn the hard way, and hopefully inspire a moment of pause about their own risk exposure tolerance.
The subject is about young workers that acknowledged the very real churn part of the ephemeral tech sector, and position themselves for an early exit strategy into adjacent careers or retirement. This is not a new idea, given vertical movement in software careers has been stagnant for decades. =3
Open desktops are like an architect that keeps moving the bathroom doors in a building under perpetual construction. The ecosystem naturally destroys itself every major release cycle, work put into quality of life improvements is lost, and code smell is so terrible people just abandon the projects after it becomes burdensome.
The mistake people make is trying to reason with people about which groups breaking changes take priority. This is why Cinnamon continues to run desktops on systems with broken GPU drivers, and installed as a fail-back GUI plan when Wayland glitches up.
People need to stop telling new users to follow a 15 step CLI tutorial to bring up a broken distribution Desktop. =3
Note, whomever floated that home-equity loan will be happy with their won profits. They win-win even if the loan defaults. Not everyone is so lucky... =3
If people live a slow-burn lifestyle, buy revenue property one can later demolish as needed, park unused cash in index funds, and split 7% between bullion & high-risk. Most folks will be fine when things eventually go sideways.
Some people like to gamble, but they usually don't talk about their mistakes. =3
LLM are great for context search, but regulatory capture and despotism will not make "AI" something it fundamentally can't become without more energy than a third of our galaxy. Picking a war with math is disappointingly absurd, but not unexpected given the obvious degenerative geriatric condition of the glorious leader. =3
Very wise, energy constraints are already feeding the hyper-scale gamblers their own hubris. =3
https://apnews.com/article/huawei-ai-chips-nvidia-superpod-t...
Take it lightly until the benchmarks drop. ymmv =3
There is a serious alternative to NVIDIA "AI" hardware dropping out of China in February 2027. There is no moat, but a whole lot of unpaid debts in the near future.
Popcorn ready =3
While fission reactors are still the cheaper initial energy choice (like a future cancer patients love of cheap cigarettes.) Small Modular reactors are a fools errand for most places, as they have all the same security, labor, and maintenance issues of full sized installations. Thus, instead of repairing one of four turbines every few years, the contaminated components must now be reprocessed or safely hidden away on mass.
Maybe fine for submarines or aircraft carriers, but for massive infrastructure it is trivial to prove it causes major long-term problems.
Anyone that has ever purchased a property with an old leaky heating-oil tank nasty surprise site-reclamation cost... will understand the initial unit price never covers the actual clean up cost. =3
Whenever you hear of labor shortages, it is almost always a euphemism for a shortage of people desperate enough to tolerate a terrible working arrangement.
The "AI" slop from staffing agencies, lead-generators, applicants, and data scrapers... have dropped the signal to noise ratio into spam number territory. =3
Xiaomi and Huawei will likely adapt their supply chains accordingly. =3
Perhaps it is time to get outside for a walk to lower stress levels. =3
If you still don't understand, than you don't understand how LLM are made.
Just because something is publicly accessible doesn't mean it is Public Domain. Having hosting people pay for the bots stolen bandwidth (or DDoS), is also theft of service under the law. =3