Joel_Mckay

Given the price for a rack of bc-250 after the crypto hype cycle, the expected value of the hardware will be around 5% to 10% of the original retail price.

Without other market influences, that is a >90% expected discount when the over-provisioned market must inevitably self-correct.

If the Market follows what Samsung/SK Hynix did to the South Korean exchange this week, than the "AI" bubble will hit harder than the dot com crash.

I like the Shrek Movie correlation theory, as they always happen just before Debt-backed investors get hit hard... And the new film is due out in 2027. =3

In general, the Embedded space is very different from multitasking Application environments. The mistake is assuming language choices people make were independent of the language specific features outside isomorphic code implementation.

Python is great in many ways, but anything with a garbage-collector or JIT has fundamental problems with parallelism and or real-time environments.

https://en.wikipedia.org/wiki/Clock_domain_crossing

https://en.wikipedia.org/wiki/Metastability_(electronics)

I would never claim to know what I am talking about, but reliance on other peoples hobby projects can inject naive/unstable dependencies into your long-term work.

This list of rules applies to most languages:

https://en.wikipedia.org/wiki/The_Power_of_10:_Rules_for_Dev...

A bit of history where these rules came from, and why they matter.

"Why Fighter Jets Ban 90% of C++ Features"

https://www.youtube.com/watch?v=Gv4sDL9Ljww

The basis of the cited 10 rules helped form many standards for reliable code:

https://www.stroustrup.com/JSF-AV-rules.pdf

Good luck, and I agree dealing with bitmap multi-language fonts is difficult without a header template generator. imagemagick can build this for you with a bit of one-time scripting, but it always eats up a lot of rom memory. =3

AI has absolutely crushed the barrier to entry for using this gear

Sure, now people can make more frequent dangerous assumptions with twice the overconfidence. =3

Pull off the C/C++ band-aid already, it will give you a clearer understanding of how these chips actually function. =3

The Esp FreeRTOS projects are a great little ecosystem, and will offer quite a bit more insight than Arduino pseudo-C++ silent-polled architecture.

In general, it depends what you want out of the ecosystems. For quickly setting up a proof of concept, an Arduino used to be good for the exhaustive documentation, examples, community, and labs.

The modern Arduino integrated product line is now similar to other locked ecosystems like FPGAs. It no longer lets students learn by burning out $10 parts they made. The only people having fun now are the folks building those trademarked modules.

Probably wise to avoid LLMs if you are a student, and check out the $6 Pi Pico 2 if interested in low-level firmware build processes (it also supports the old Arduino IDE if you get stuck.)

Best of luck =3

Lab:

https://learn.adafruit.com/using-raspberry-pi-pico-pico-2-wi...

Pico 2:

https://www.mouser.com/en/new/raspberry-pi/raspberry-pi-pico...

https://www.mouser.com/en/ProductDetail/Raspberry-Pi/SC1632?...

Breadboard and common components to build and blink LED lights (set everything at 3.3v to start exploring):

https://www.amazon.com/EL-CK-002-Electronic-Breadboard-Capac...

pick and place board assembly means through-hole components are going the way of the dodo

The factories willing to do SMD mixed with through-hole high-power electronics is shrinking. However, package constrained power dissipation limits on SMD technology is real.

Depends on what your firm builds. =3

Distilling models using more advanced LLM is not a new phenomena. It is a cost effective strategy in a highly competitive emerging field.

Also, the Hidden-Agent problem exists in every model, and is a persistent tangible risk independent of whatever team people cheer for at the games. Let us remember, every LLM nuked all of humanity 92% of the time in simulated war games. =3

"AI" responses are silly, and optimal 24 core count efficiency premise in Desktop applications offer diminishing returns on a highly concurrent 32/64/192 core Epyc line of OoO chips...

How many strings does a Bass play with in water? =3

Why cross-entropy follows a curve is fairly trivial to understand:

https://www.youtube.com/watch?v=GlYgs6v2YfU

In general, motion planning multi-axis systems is NP-Hard. Have a great day =3

Indeed, people are still bad a parallelism today, and most compilers still suck at reliably unrolling abstracted concurrent source intended functionality.

Very few modern languages handle parallel scaling gracefully, and bodged on CUDA still isn't great either. As Moore's law ends, people have to reevaluate how they approach traditionally monolithic architectural design. =3

Indeed, the post does not conflate the two... look closer. =3

Indeed, my point was not all chips do well with Desktop application loads.

People were bad at handling parallelism with ia64, and still have problems today on better amd platforms with less janky compilers.

The fact an $800 chip still can beat a $14000 chip at some tasks probably should tell people something about concurrency scaling overhead. =3

We don’t know what it would look like if Apple had chosen something that ended up like Itanium

Apple made a few mistakes, but mostly by trying to compete with doomed hype markets like AR/VR.

Some also ponder what the ecosystem would look like today if the Windows NT kernel had stayed on RISC like initially planned.

The "What if __ ?" universe are fun to imagine, but ultimately less important than the "What now?" universe we live in. =3

My point was most modern software developers still do not understand systems at an architectural level. I don't buy the VLIW paradigm was inherently inferior argument, but the code written for it with mystery failure modes at the time was a mess.

And ia64 was only around 23% slower and several times more costly than amd64 options at the time. Both toads had their warts, but one was actually usable by mere mortals.

Its true just about everything either ran slow on not at all on ia64... Thankfully, we were still running Sparc based platforms at that time, as the Intel clown-show still looked like more work. But I do empathize with the ia64 trauma, as the roll out had a lot of collateral damage in some firms as people started jumping ship early to avoid accountability. =3

Also may be irrelevant for people at some jobs, and cause cognitive skill losses over time.

https://www.youtube.com/watch?v=axOcn--n_lM

https://www.anthropic.com/research/AI-assistance-coding-skil...

We should remember LLM do have legitimate use-cases like search, as we enter the "Trough of disillusionment" in the hype cycle. =3

https://en.wikipedia.org/wiki/Gartner_hype_cycle

Agreed, everyone has a pet architecture that fascinates them, but if its so good only 5 people can code for the target... it is e-waste within a year.

Apple is an exception as it has always had a walled-garden ecosystem with the OS, so can force shifts in architectures unlike most companies. The M3/M4 Pro series with unified GPUs is probably the best design on the consumer market right now, but people are not leveraging it as much as they would have in other ecosystems.

Have a wonderful day =3

We agree it probably would have still failed, but mostly it was the legacy code-motion compatibility/performance issues that were practically inescapable without refactoring millions of lines of code.

gcc maintained the ia64 target a long time for unclear reasons, but it was also still inefficient on other platforms. The FOSS compiler worked, but that was its only performance metric that counted for many users. =3

Not sure why you think the Intel compilers were a myth, as they are still around working far better than gcc in many use-cases:

"An Overview of the Intel® IA-64 Compiler"

https://webdocs.cs.ualberta.ca/~amaral/courses/605/papers/In...

To be fair, most of ia64 performance issues were in the more complicated compilers. The chip itself had potential, but essentially abandoned the DOS/Windows market inertia. =3

The ia64 more complex compiler ecosystem made it difficult or impossible to port performant programs to the platform. Amd64 allowed people to keep building and running most programs with little understanding of what was happening in the architectural layers.

Even the 800lb guerilla Intel had to acknowledge amd64 was the next step, or it could end up a niche use chip like PowerPC.

The moral of the story is usability matters with consumer chip architecture. =3

Nice, sounds like a lot of fun. A few years from now one may have 2 businesses, as solving the niche issues is how small business happens. =3

We found most of this devs work was popular with artists. May be worth a look.

https://bartoszstyperek.gumroad.com/

LLM are sometimes good at organizing the awful documentation for Blenders ever permuting API, but will not help you get better creatively. If you are interested in geometry nodes, than this channel offers a lot =3

https://www.youtube.com/@TheDucky3D/videos

Indeed, geometry nodes are revolutionary in some ways, and volumetric displacement brushes are magic.

The issue is Blender has been in perpetual Beta for decades. Every update something breaks, moves, permutes, or goes missing. It is chaos incarnate in some ways. =3

Mostly Fun... sometimes replacing commercial software workflows that were already a disaster for the creatives. =3

There is no such thing as "free" unless the original author released files under CC0 or Public domain. The rest of the ecosystem is a legal minefield even for kids kit-bashing a project for fun.

We spent around $3k conducting a survey of the Blender ecosystem for a game dev project about 2 years ago, and continue to monitor the situation every update. A rather sobering experience in some ways.

Would recommend Poly Haven as the plugin is not required to access their free assets, but it is usually a good idea to support people that make things easier for everyone. =3

https://polyhaven.com/plugins/blender

Every use-case differs. I like Blender even when they break my hobby projects every update.

OpenScad and FreeCad also works for 3D cad files. However, many people that own 3D printers will never sculpt, re-mesh, or cad up their own objects. =3

https://krita.org/en/

It is also free. =3

Erlang ran multi-billion dollar market cap telecom companies for decades. Elixir just made Erlang easier to learn, and yes they replace a lot of cluster bodges at scale.

Does rapid cross-platform deployments with single code-base ease, and with Phoenix channels is very economical =3

https://quasar.dev/introduction-to-quasar/

It certainly sounds similar, but seems more nimble with phonetic pronunciation in the demo.

Having it run on a pico would be pretty impressive =3

http://cmuflite.org/

https://github.com/festvox/flite

Industry leading Interactive Voice Response systems have become very good at filling in ambiguous information from context, and modulating pronunciation to Ape emotional information.

However, being able to interact with these natural language systems in uncontrolled settings is still a fools errand. For STT, there is also regional dialect, slang, and individual differences.

Witnessing blind users hit unrecognizable reading-speeds on old Gordon 8 TTS systems was surprising. I learned people adapt to imperfect systems pretty quickly. =3

True, but a blacklisted $5k GPU seems like a bit of a risk. =3