SoylentNews is people
SoylentNews
MBed OS and platform are shutting down in 2026, although rumor has it almost all of the devs have already been downsized.
https://os.mbed.com/blog/entry/Important-Update-on-Mbed/
A couple of possible discussion points from the perspective of someone who used it for STM32:
It was one of those FOSS-but-not-really products that was completely corporate controlled and funded and written, but under a FOSS license. It never really gained any traction outside corporate. There is a winner-take-all mentality in microcontroller RTOS... why use Mbed if Zephyr supports 10x as much "stuff" out of the box? Also, given the primary source of funding, it really only practically functioned on ARM processors. Pragmatically it seems multiplatform RTOS are the only ones that survive long-term, single platform seems always doomed, a bit different than the desktop/laptop/phone market.
There was something of a product-tying thing going on with Pelion IoT cloud platform, which used to be free, but the free tier disappeared. It was pretty awesome for hobbyist use until they intentionally got rid of the hobbyists, presumably to "save money". However this seems to be a common pattern for decades, the devs who influence million dollar contracts during the day want to play with pirated/free versions at home at night, so arguably Pelion and thus Mbed shot themselves in their own foot.
I wonder how much C19 killed Mbed a couple years later. After STM32 procs and ARM microcontrollers were unobtainable for couple of years, there was no way to get hardware to run Mbed.
It was a bit memory-hungry; IIRC by the time you got a full IoT platform with auto-updates and telemetry over WiFi working on commodity dev board hardware, you were out of either flash, ram, or both so you couldn't run your app.
I have happy memories of being introduced to LwM2M protocol; it was an interesting innovation on MQTT but a little too "organized" for widespread use. Take MQTT and "compress" by turning all common (and uncommon) nouns and verbs into integers; kind of like the old Apollo spacecraft computer, kind of like a fixed compression standard.
A final interesting discussion point is tool manufacturers going out of business is a pretty strong signal the bubble is over. The permanent solution to "The S in IoT stands for security" may very well be the IoT industry drying up and blowing away, and this shutdown is a sign of the start of the end.
Anyone else have fond memories of MbedOS? I thought it was pretty awesome back in the day, although I switched to Zephyr years ago. Other contemporary microcontroller or IoT comments?
(Score: 2) by acid andy on Tuesday July 16 2024, @11:26PM (4 children)
I just want to stop here and reflect on the fact that we are casually talking about bots passing the Turing test like it was never a big deal. Strange times we live in.
"rancid randy has a dialogue with herself[...] Somebody help him!" -- Anonymous Coward.
(Score: 3, Interesting) by pTamok on Wednesday July 17 2024, @09:38AM (1 child)
"...we are casually talking about bots passing the Turing test like it was never a big deal."
Nope.
Some people are casually taking about bots passing what they think is the 'Turing Test', without knowing what they are talking about.
It's the Imitation Game, and does not involve a person deciding if entity they are talking to is artificial or not*. It also happens to be a statistical measurement over many 'runs'. Turing's original paper is available without charge online. It is worth reading. Most social network commentators 'discussing' AI definitely have not read it, or if they have, have not understood it.
Furthermore, intelligent robots/computers/non-biological things have been a trope in Science Fiction for a Very Long Time. Hoyle's 'The Black Cloud' from 1957 was more intelligent than humans. Asimov's robots with 'positronic brains' ate also quite well known. Everyone else is just catching up.
*The key sentences in Turing's paper, which many people ignore, are:
The object of the exercise is to determine whether A is a man and B is a woman, or B is a man and A is a woman. It is not to determine is A is a man or machine. Effectively, the game measures how well a machine can simulate a man pretending to be a woman (for the purposes of the game), and convince the interrogator (erroneously) that the women is more likely to be the man. It seems to be too subtle a point for most readers and commentators.
(Score: 2) by acid andy on Wednesday July 17 2024, @11:46AM
OK I can see what he was getting at but imagine the controversy if such a methodology was used today. It relies on the theory that men and women talk differently to a degree that can be observed beyond random variation. In Turing's day I am sure the social pressure to conform to very distinct gender roles would create very different dialog between men and women on average. Today in the west I am sure the very idea would cause outrage and long arguments.
Anyway, I've learned something about the Turing test, so thanks for that.
"rancid randy has a dialogue with herself[...] Somebody help him!" -- Anonymous Coward.
(Score: 2) by VLM on Wednesday July 17 2024, @01:12PM (1 child)
A lot of social media posting is like prayer; send it out into the void for a possible updoot, or not; back and forth discussion like on Soylent News is relatively rare.
(Score: 2) by acid andy on Wednesday July 17 2024, @01:57PM
I know and it's quite depressing. Even on reddit discussions rarely seem to last more than a couple days. Hell, even stack overflow will have a limited number of replies in a disussion although I appreciate that is quite heavily controlled to steer things towards a well defined question with a well defined answer.
Longer attention spans are needed for people to really properly think about a topic and analyze it in depth. The human race is so fucked.
"rancid randy has a dialogue with herself[...] Somebody help him!" -- Anonymous Coward.