SmartSDR v4.0.1 and the SmartSDR v4.0.1 Release Notes
SmartSDR v3.10.15 and the SmartSDR v3.10.15 Release Notes
The latest 4O3A Genius Product Software and Firmware
Need technical support from FlexRadio? It's as simple as Creating a HelpDesk ticket.
Asked Google AI if the new noise reduction features would work with 6000 radios
And the answer is:
AI Overview
Yes, FlexRadio's new noise reduction features, including AI noise reduction, are designed to work with the 6000 series radios. These new features are part of the SmartSDR v4.0.1 software and firmware update, which is compatible with both the 6000 series and other Flex radios like the Aurora and FLEX-8000.
Software Compatibility: The new features are included in the SmartSDR v4.0.1 update, which supports the 6000 series radios.
AI Noise Reduction: The release notes explicitly mention "Noise Mitigation" and a new "AI noise reduction" feature, confirming it is available for compatible radios.
Affected Radios: The update is for the Aurora, FLEX-8000, and FLEX-6000 series of radios, as well as the Maestro.
Update Details: You can update to the new features by installing SmartSDR v4.0.1 or later, which is available for free as part of the basic SmartSDR package.
Comments
-
I know that this answer is full of holes. But, it shows how you can not trust everything on the internet even if it is given by so called, AI.
James
WD5GWY
1 -
Just to note, Google AI doesn't have anywhere near the deep thinking power to answer technical questions with deep specifics. I know you were not looking for detail. Your question was general on purpose. And so general answer. Now, for others, I would just make a note here that if you happen to use ChatGPT 5 (in the deep thinking mode) or Grok 4 (with pro analytics), that is a whole different level of knowledge generation.. It actually knows or researches very deeply, provided it is driven by some well crafted and detailed prompts, challenges and guardrails. And does it at lighting speed. Which you have discovered, This I am sure you know. It is all in the prompting design. The reason people are developing training courses on prompting. And some are making careers of prompting. AI Prompt Engineers. Big $$$. hi hi Now, for Flex radio technical topics, the results can be a pretty mind blowing and provide comprehensive answers that would satisfy most. It is almost like have a team of RF / DSP engineers at your disposal. Pretty Wild. 73 Dan WA1QZX
1 -
Thanks for the offer. I have seen your results on the new features. I posted this mostly in jest. And in hopes to lighten the mood around here over the different misunderstandings of what new features in SmartSDR will or will not, work with the 6000 series radios. Google's AI in their search feature can return some odd results. And that is what I was trying to convey. Nothing more.
James
WD5GWY
0 -
Thanks James. I got your intent of lightening things up now. Good idea. Been pretty confusing around here lately. Maybe that's the right word: Confusing? I edited my message to soften it up, I just really want to make a note of using more advanced AI agents and focus efforts on prompting design. If you would not mind re-reading my edited message, if it came though OK, it would make more sense.
1 -
Looks good to me.
James
WD5GWY
0 -
Here is the compatibility matrix:
https://us.v-cdn.net/6032377/uploads/7GM5U9WB5HIA/screenshot-2025-10-30-at-2-07-12-e2-80-afpm.png
73
1 -
AI on FlexRadio products is about 70% accurate. You need to figure out which 30% is wrong and you need to tell it not to make assumptions.
1 -
Depends upon which "AI" you use and how well your prompting is designed. ChatGPT 5 (deep thinking mode) and Grok 4 (pro analytics) provide accuracy well above 95+% which was not true only 6 months ago. Getting better everyday. It's all in the prompt design phase.
0 -
Dan, maybe for your world. We’re not seeing quite the same thing for FlexRadio direct questions. The technology is impressive, but it’s not perfect—and the inaccuracies we’re seeing from AI responses directed at FlexRadio topics have actually increased our post-sales support load. Customers often take AI-generated answers at face value, which then leads to confusion or misconfiguration issues that we have to resolve manually. It is great at making wrong assumptions.
It’s not that AI is bad—it’s improving rapidly—but for technical or product-specific queries like ours, the margin of error is still high enough to cause operational friction. We end up spending time clarifying and correcting these AI-driven misunderstandings on a customer-by-customer basis. Until these models have tighter alignment with verified product data, we’ll likely continue seeing this gap between general AI confidence and actual accuracy for niche domains like ours.
2 -
Mike, I disagree with you in a sense. It depends, you can use AI poorly or in an exceptional way. If you don't put the time and effort into detailed prompt design ( w/ challenges, guardrails and citations) you will not have the best results. However, if you do prompt design well, you will get very very accurate, detailed and beneficial results that you can put your full weight on. The only thing it can't give you is unknowable proprietary internal information that is not available outside of Flex. But, that is the only limitation, at this point. Inaccuracies (hallucinations) are basically a thing of the past today. This was not the case say only 6 months ago. But, again it is all in the prompt design. You are basically beam steering with narrow angles into the niche technical domain at a levels of knowledge, sophistication and nuance that is actually beyond the capabilities of most engineering teams within the same niche (unless they are using AI themselves, which most all do now, or should be best practice). If you're designing say an new advanced RF or Audio filter for example, you're using AI today to design it. No question. Anyway, my 2 cents. 73 Dan WA1QZX
0 -
Dan, I am curious, do you work with a company that produces AI systems? Not being critical or anything like that. But, you seem to be all in on using AI for many things. I think used as another tool for things like software development, it could be fine. But, I still feel human imagination can offer things that no amount of artificial intelligence could ever match. I feel like human creativity can go places that a machine can't begin to imagine without human input guiding it in some manner.
So, just a thought on my part.
James
WD5GWY
0
Leave a Comment
Categories
- All Categories
- 368 Community Topics
- 2.1K New Ideas
- 625 The Flea Market
- 8.1K Software
- 74 SmartSDR+
- 6.4K SmartSDR for Windows
- 182 SmartSDR for Maestro and M models
- 421 SmartSDR for Mac
- 270 SmartSDR for iOS
- 255 SmartSDR CAT
- 189 DAX
- 381 SmartSDR API
- 9.3K Radios and Accessories
- 33 Aurora
- 238 FLEX-8000 Signature Series
- 7.2K FLEX-6000 Signature Series
- 937 Maestro
- 54 FlexControl
- 863 FLEX Series (Legacy) Radios
- 913 Genius Products
- 460 Power Genius XL Amplifier
- 331 Tuner Genius XL
- 122 Antenna Genius
- 293 Shack Infrastructure
- 206 Networking
- 449 Remote Operation (SmartLink)
- 144 Contesting
- 773 Peripherals & Station Integration
- 139 Amateur Radio Interests
- 991 Third-Party Software
