@world101 i just manage to have a partial sucess using lastest firmware V1.4.5.1, here my findings
a) My Blockly app (which worked last year) contained a loopback test i.e. both a publish and substribe MQTT within the same application (it makes sense just for testing) but that is the very reason i was having problems (app freeze before first statement and "connect fail displayed in red), if i leave the publish statement only then my app does not freeze and after changes on the publish topic including username/feeds/topic (using adafruit server io.adafruit.com and my new credentials) and the Adafruit dashboard correctly display the published value (pressure, emulated with the rotary potenciometer, my ref: Environment_IoT34E_test3_Adafruit_Dashboard_example.png) .
b) Despite successful test the behaviour is still UNRELIABLE while booting, still gettin the 'connect fail' message in red after boot ~ every second attempt BUT the application does not freeze anymore and continue working as intended and publishing the topic data as explained above.
Again, thank you very much for your "slight' hint about Adafruit difference with cloudmqtt.com.

@world101 thank you for your reply, i tried to look for examples that you mentioned and i saw one of your post where the MQTT publish and subscribe topics have to be preceded by the path at the adafruit server ( io.adafruit.com ) username/feeds/topic and i did the same...but that does not solve my very first problem, my Blockly app (on Core nor on Stick-C) never get to execute the very first statement after Startup and the app is frozen with overlapped error message in red "connect fail"...i did ad a waiting loop for the wifi to be ready....i may have forgotten to mention that i am in Shanghai/China and maybe the internet sucks as usual and the M5Stack OS timeout too early? ...just speculating..but this issue cannot be related to MQTT....the app should execute the rest of the code at least but it is not the case, it is just frozen after showing initial screen (drawn manually in Blockly)....

a) I did manage to program simple UIFlow / Blockly app into the ATOM Matrix with UIFLow V1.4.5 but i failed to do so consistently and non even one i managed to complete my intended simple application because UIFlow half-froze (both Browswer version and Desktop one) and e.g. in that state i clicked on the Python tab and was not possible anymore to go back to the Blockly tab, nor possible to save the app, so i lost all my work , multiple times!
b) The internal webserver intended for wifi SSID/password configuration to get the API Key as explained in the documentation NEVER showed up and therefore it was necessary to use a console (PuTTY in my case attached to the virtual COM port) to be able to get to the API Key; such simple product is intended to be used by children but requires software engineering know-how to deal with all that "low level" stuff; not good!
c) I also tested the ATOM Matrix with Arduino IDE 1.8.10 and tried some of the about 4 examples given and the behavior was also extremely unreliable, more than one time while or after app firmware download the ATOM Matrix was in an undefined state , nothing shown on the LED Matrix and it was necessary to erase and re-flash the firmware with the Burner. It was not possible to run all examples.

A) Issue A: M5Stack IoT tutorial and example are based on cloudmqtt.com; most of user like me just use “Cute Cat” Free MQTT broker which it is "out of stock" i.e. cloudmqtt.com does not offer the free tear anymore. So i kindly ask M5Stack to test and document with a tutorial other options.
B) Adam Bryant from M5Stack FB Group suggested to try Adafruit MQQTT broker, which i did; nevertheless immediately after my Blockly app boot, when trying to publish the first value (of pressure) the Adafruit Dashboard reports "MQTT ERROR: rejected, not a valid topic format"...the application i am trying did work perfectly last year with cloudmqtt.com “Cute Cat” Free MQTT broker though , i just changed the credentials during MQQT initialization. Therefore it looks to me that there is a MQTT protocol compatibility issues between Adafruit IoT platform/MQTT broker and M5Stack IoT / MQTT / Blockly/micro-Python compatibility issues? ....
C) For a discussion and documentation with pictures of all the problems i am facing with Stick-C/Core/UIFlow V1.4.5 related to WiFi and IoT please check the thread of discussion on the FB M5Stack group from March 15th onward....
Thank you very much!

I bought a first 20cm LED strip (containing 29 LEDs) ...they work fine...but the problem is that the LED strip has two male connectors and therefore cannot be connected to the Core e.g. Port A directly but it needs an extenstion I2C adapter cable...and if you want to connect multiple strips in series that would require one of such adapter cable for every additional LED strip segment....that does not make sense to me and visually it would not be nice....why you guys didn't do the strip with one female and one male connector??

@fonix232 i dont understand what you mean...this issue appeared e.g. this morning when the CPU had been off over-night, i wanted to power it up (without plugin in USB) and then the power up would not work and the audible noise is emmited...and as i reported into another bug the CPU would not power up.....Question, i got to USB cables for the M5Go, the communication only works with the very short black one but not with the small white "docking station"....is that normal? i wish i could work via USB via the USB docking station but it seems to be that that is only to charge the internal battery??

@ajb2k3 ...interesting....i haven't seen the screw kit on the M5Stack official Taobao store , i hope it comes out soon. I bought the Servo module there, M5Stack official Taobao store.
Regarding the use of MicroPhyton, which IDE is recommended? In any case as i mention i plan to organize programming and robotic courses for children, youngsters and mainly non-tech adults so i have to stick to Blockly.... i really hope the M5Stack R&D team (i tough you worked for them) manage to stabilize the platform sooner than later. Thank you! JC

This bug on UiFlow-Desktop V1.2.3 (Windows 10 64bit via USB - target M5Go) appears from the very first day i used it more than a week ago. Very often if you have added multiple text labels to the virtual M5 screen on the UiFlow-Desktop you are not able to re-select them individually to modify its atributes i.e. the label number, text content nor font type. Let's say i have five labels on the virtual screen then on anyone i click it just show me the very same label information (not sure which one, probably just the last one i added just before the bug showed up). This is extremely annoying as the only workaround is to define the proper text content programatically BUT for example the font size cannot be set programatically so the application just looks ugly and the default font is too small. Please fix this! (NOTE: i cannot say if the bug is also present on the UI-Flow browser version because the connection via internet is extremely unreliable here in China and it almost never connect).

@ajb2k3 this morning i reserved the servo kit and i successfully installed/stacked ...although i am not able to use the original screws to properly stack all 3 layers together...shouldn't additional stack kits come with four longer screws to fix the whole stack??...anyhow that is not the point of this reply; what i wanted to mention is that after programming the servo application almost identically as in the tutorial (https://m5stack.com/video) i was actually able to (using UIFlow-Desktop V1.2.3) download and run the target (M5Go) i.e. while changing the application program, without the need of manual reset as i had reported above. Maybe the fact that i reopen it and reconnected/pressed the stack layers together changed something?...it remains a mistery.....the lack of predictability /reliability of that behavior and the one of the manual/power on reset are very annoying....i was thinking to use the M5Stack HW/SW to offer computer science and robotics to children/teenagers and nonn-tech adults but with such unreliable behavior of the M5Stack HW/SW it is really difficult to go for it; it would result in frustration from the students and maybe on a total failure of the training attempt.

@fonix232 i have the same problem of the noise when pressing the reset & power on button ( i mentioned in one of my separated posts)...and worst...it is very unpredictable when the Core (my main kit is the M5Go) will power up...this morning i have to try over 10 times...and it seems it will only power up if connected to USB. You said you had the noise issue with the first kit and with the second kit ....i have it with my own kit, so @ajb2k3 this issue should not be deemed [Solved] or a new bug/ticket should be open!.