MDhere
Top 20
How many shares in your portfolio?I want a brainchip shirt.

How many shares in your portfolio?I want a brainchip shirt.
You can have mine. Hardly used. I only wore it if the share price was over $1. So, it's like brand new.I want a brainchip shirt.
Yes I start MondayHas the social media position been filled yet anyone know ?.
Yes, money to be made if they produced a "BRAINCHIP DISGRUNTLED SHAREHOLDER" oneI want a brainchip shirt.
Fuck Off.!!RECOVER YOUR SCAMMED CRYPTO FUNDS Call iFORCE HACKER RECOVERY
iFORCE HACKER RECOVERY rescued my shyness and inexperienced with cryptocurrency, I made a terrible mistake. Influenced by my in-law, who seemed confident about investing, I put over £77,000 my savings for a new home into what turned out to be a scam. When the promised returns never came, I realized I’d been swindled. Devastated, I searched for help and found reviews of iFORCE HACKER RECOVERY. After speaking with someone they’d helped, I reached out. Following their guidance, I was amazed to recover most of my funds. They turned my despair into hope. I'm truly grateful.
Contact iForce Hacker Recovery:
Email: iforcehk@consultant.com
WhatsApp: +1 (240) 803-3706
Website: iforcehackers.wixsite.com View attachment 88043
That's top secretHow many shares in your portfolio?![]()
What if she signs a NDA?That's top secret
And here is a third one showing Kurt Manninen with the LLM demo:
I wish they would not use the exact same question every single time (“What is a point of sale system?”).
View attachment 88047
PAll this talk about T-shirts has got me spiralling back to that moment when Rob Telson declared that Ken the robot needed a companion - a robotic dog named Higgins, of all things.
Did he consult a psychic who secretly foreshadowed the Chelpis quadruped dog-bot project?
Why the name Higgins?
What does it mean?
What is the meaning of life?
Is there only one universe or is this a simulation?
If there's a multiverse, can we collect licensing revenue from alternate dimensions too?
I often find myself waking up at about 4am trying to answer this conundrum.
Should I consult the new IR team about it?
Shout out to Sean and Antonio. Thanks for all that you have done for Australian investors.
Apart from radar as an obvious choice for AKIDA I think Lockheed-Martin would also look at cybersecurity to protect radar and missile systems from interference.View attachment 88056
My Question: What would be the most logical areas of contribution be, if BrainChip were to be involved in this contract/program?
ChatGPT Response: If BrainChip were to be involved in Lockheed Martin’s Aegis Ballistic Missile Defense (BMD) upgrade program under this new $2.97 billion contract, the most logical area of contribution would be in advanced radar signal processing and sensor fusion — areas where their Akida neuromorphic processor and TENNs (Temporal Event-Based Neural Networks) can offer unique advantages.
Specific Components Where BrainChip Could Contribute:
1. Radar Signal Processing (e.g., SPY-1, SPY-6 radars)
- Aegis relies heavily on radar (active electronically scanned arrays) for missile detection, tracking, and intercept guidance.
- Akida could be used to pre-process radar dataat the sensor edge with:
- Low latency and low power
- Micro-Doppler signature recognition (detecting, classifying, and predicting incoming missile threats)
- Improved discrimination between clutter and real targets
- This would reduce the processing burden on main mission computers and improve real-time threat responsiveness.
2. Sensor Fusion / Multi-modal Data Integration
- Aegis systems integrate data from radars, sonar, EO/IR sensors, and satellite feeds.
- Akida's event-based processing could help:
- Fuse diverse sensor inputs more efficiently
- Extract high-confidence signals from noisy, complex environments
- Maintain performance even in GPS-denied or jammed conditions
3. Missile Interceptor Guidance Systems (e.g., Glide Phase Interceptor - GPI)
- Neuromorphic processors like Akida are well-suited for embedded use in small, SWaP-constrained platforms like interceptors or kill vehicles.
- They could help with:
- Terminal phase threat discrimination
- Real-time object tracking and targeting adjustments
4. Electronic Warfare (EW) and Threat Classification
- As Raytheon and Lockheed both expand into cognitive EW, BrainChip could support:
- Autonomous jamming detection and classification
- Edge decision-making in dynamic threat environments
Why Neuromorphic vs Traditional AI?
Traditional AI (e.g., CNNs running on GPUs) is:
- Power-hungry
- Latency-prone
- Often not suitable for real-time edge decision-making in missiles, drones, or battlefield systems
BrainChip's Akida and TENNs offer:
- Real-time response
- Ultra-low power use (critical for embedded platforms)
- On-device learning and classification without needing cloud uplinks
Summary:
BrainChip would most likely contribute in areas such as:
- Edge radar signal processing
- Micro-Doppler threat discrimination
- Real-time, low-latency decision support for missile intercept systems
- Sensor fusion and cognitive electronic warfare
Their role would likely be as a hardware-level enabler, accelerating and enhancing AI inference and classification functions on the edge, particularly in areas where every millisecond and milliwatt counts.
"View attachment 88056
My Question: What would be the most logical areas of contribution be, if BrainChip were to be involved in this contract/program?
ChatGPT Response: If BrainChip were to be involved in Lockheed Martin’s Aegis Ballistic Missile Defense (BMD) upgrade program under this new $2.97 billion contract, the most logical area of contribution would be in advanced radar signal processing and sensor fusion — areas where their Akida neuromorphic processor and TENNs (Temporal Event-Based Neural Networks) can offer unique advantages.
Specific Components Where BrainChip Could Contribute:
1. Radar Signal Processing (e.g., SPY-1, SPY-6 radars)
- Aegis relies heavily on radar (active electronically scanned arrays) for missile detection, tracking, and intercept guidance.
- Akida could be used to pre-process radar dataat the sensor edge with:
- Low latency and low power
- Micro-Doppler signature recognition (detecting, classifying, and predicting incoming missile threats)
- Improved discrimination between clutter and real targets
- This would reduce the processing burden on main mission computers and improve real-time threat responsiveness.
2. Sensor Fusion / Multi-modal Data Integration
- Aegis systems integrate data from radars, sonar, EO/IR sensors, and satellite feeds.
- Akida's event-based processing could help:
- Fuse diverse sensor inputs more efficiently
- Extract high-confidence signals from noisy, complex environments
- Maintain performance even in GPS-denied or jammed conditions
3. Missile Interceptor Guidance Systems (e.g., Glide Phase Interceptor - GPI)
- Neuromorphic processors like Akida are well-suited for embedded use in small, SWaP-constrained platforms like interceptors or kill vehicles.
- They could help with:
- Terminal phase threat discrimination
- Real-time object tracking and targeting adjustments
4. Electronic Warfare (EW) and Threat Classification
- As Raytheon and Lockheed both expand into cognitive EW, BrainChip could support:
- Autonomous jamming detection and classification
- Edge decision-making in dynamic threat environments
Why Neuromorphic vs Traditional AI?
Traditional AI (e.g., CNNs running on GPUs) is:
- Power-hungry
- Latency-prone
- Often not suitable for real-time edge decision-making in missiles, drones, or battlefield systems
BrainChip's Akida and TENNs offer:
- Real-time response
- Ultra-low power use (critical for embedded platforms)
- On-device learning and classification without needing cloud uplinks
Summary:
BrainChip would most likely contribute in areas such as:
- Edge radar signal processing
- Micro-Doppler threat discrimination
- Real-time, low-latency decision support for missile intercept systems
- Sensor fusion and cognitive electronic warfare
Their role would likely be as a hardware-level enabler, accelerating and enhancing AI inference and classification functions on the edge, particularly in areas where every millisecond and milliwatt counts.
Also AKIDA's chip dynamic learning is a huge step up from pre programmed static existing systems. AKIDA will adjust to changes in threats in real time which is critical for defence systems.View attachment 88056
My Question: What would be the most logical areas of contribution be, if BrainChip were to be involved in this contract/program?
ChatGPT Response: If BrainChip were to be involved in Lockheed Martin’s Aegis Ballistic Missile Defense (BMD) upgrade program under this new $2.97 billion contract, the most logical area of contribution would be in advanced radar signal processing and sensor fusion — areas where their Akida neuromorphic processor and TENNs (Temporal Event-Based Neural Networks) can offer unique advantages.
Specific Components Where BrainChip Could Contribute:
1. Radar Signal Processing (e.g., SPY-1, SPY-6 radars)
- Aegis relies heavily on radar (active electronically scanned arrays) for missile detection, tracking, and intercept guidance.
- Akida could be used to pre-process radar dataat the sensor edge with:
- Low latency and low power
- Micro-Doppler signature recognition (detecting, classifying, and predicting incoming missile threats)
- Improved discrimination between clutter and real targets
- This would reduce the processing burden on main mission computers and improve real-time threat responsiveness.
2. Sensor Fusion / Multi-modal Data Integration
- Aegis systems integrate data from radars, sonar, EO/IR sensors, and satellite feeds.
- Akida's event-based processing could help:
- Fuse diverse sensor inputs more efficiently
- Extract high-confidence signals from noisy, complex environments
- Maintain performance even in GPS-denied or jammed conditions
3. Missile Interceptor Guidance Systems (e.g., Glide Phase Interceptor - GPI)
- Neuromorphic processors like Akida are well-suited for embedded use in small, SWaP-constrained platforms like interceptors or kill vehicles.
- They could help with:
- Terminal phase threat discrimination
- Real-time object tracking and targeting adjustments
4. Electronic Warfare (EW) and Threat Classification
- As Raytheon and Lockheed both expand into cognitive EW, BrainChip could support:
- Autonomous jamming detection and classification
- Edge decision-making in dynamic threat environments
Why Neuromorphic vs Traditional AI?
Traditional AI (e.g., CNNs running on GPUs) is:
- Power-hungry
- Latency-prone
- Often not suitable for real-time edge decision-making in missiles, drones, or battlefield systems
BrainChip's Akida and TENNs offer:
- Real-time response
- Ultra-low power use (critical for embedded platforms)
- On-device learning and classification without needing cloud uplinks
Summary:
BrainChip would most likely contribute in areas such as:
- Edge radar signal processing
- Micro-Doppler threat discrimination
- Real-time, low-latency decision support for missile intercept systems
- Sensor fusion and cognitive electronic warfare
Their role would likely be as a hardware-level enabler, accelerating and enhancing AI inference and classification functions on the edge, particularly in areas where every millisecond and milliwatt counts.
there is temperature in ai models settings. when you set to 0 it will always give you the same answer. when you set to bigger number it will always give you different answersThey may have asked the same question , but the responses are totally different . Has Kurt’s model been fine tuned specifically for retail purposes ?
View attachment 88054
View attachment 88055
Oh no!!!"
"
- Aegis systems integrate data from radars, sonar, EO/IR sensors, and satellite feeds.
- sonar for hypersonics - it's not "here it comes!", it's "what was that?" ... and then "Duck!"