Sign Up | LinkedIn
500 million+ members | Manage your professional identity. Build and engage with your professional network. Access knowledge, insights and opportunities.
Please read for comments, here's a couple.
I assume it's the IBM software called Symphony that's being talked about.![]()
Sign Up | LinkedIn
500 million+ members | Manage your professional identity. Build and engage with your professional network. Access knowledge, insights and opportunities.www.linkedin.com
Please read for comments, here's a couple.
View attachment 94766 View attachment 94767
![]()
Sign Up | LinkedIn
500 million+ members | Manage your professional identity. Build and engage with your professional network. Access knowledge, insights and opportunities.www.linkedin.com
Please read for comments, here's a couple.
View attachment 94766 View attachment 94767
Talk about morphic resonance. I just posted about the change from a packet-switched mesh to a silicon-switched mesh in Akida 3, and here's JT talking about the packet overhead. The navigation/timing data takes many more bits than the spike information ... and that means power and time. So Akida 3 is much more efficient than Akida 2.Digital or analog ?
Maybe Olivier can clear it up for us ……
View attachment 94750
View attachment 94751
Thank you Olivier for clearing this up …….
Over to you JT ,
View attachment 94756
View attachment 94755
And while we are jibbering along about time and timestamps , I think this is new from Prophesee and OpenMV , it could also explain why Brainchip staff keep liking openMV on LinkedIn .
View attachment 94757
View attachment 94758
But they can't bury you if your heart's still beating ...Interesting
The "no-charge" future is officially in production. China’s Betavolt has started mass-producing the world’s first coin-sized nuclear battery. Using nickel-63 and diamond semiconductors, the BV100 provides 50 years of continuous power without a single charge. It’s fireproof, radiation-safe, and eventually decays into harmless copper. While it’s currently designed for sensors and medical implants (imagine a pacemaker that never needs a second surgery!), the 1-watt version arriving this year is the first step toward a world where your devices outlive you.
But they can't bury you if your heart's still beating ...
I don’t know but he was caught on cctv the other day. I’m guessing he just failed in getting his 1st signed contract again.Sean when shareholders ask him what he is doing for almost 5 years
![]()
I don’t think they intend to use Loihi research chips ……
View attachment 94752
Quarterly report ….. 10,000 , such an odd even number .
View attachment 94753
Clicked on the Darpa link ,
View attachment 94754
Talk about morphic resonance. I just posted about the change from a packet-switched mesh to a silicon-switched mesh in Akida 3, and here's JT talking about the packet overhead. The navigation/timing data takes many more bits than the spike information ... and that means power and time. So Akida 3 is much more efficient than Akida 2.
Dodgy , also add this to your collection of one day I might need to take into consideration ,
View attachment 94776
Which binned do you think Olivier is referring to ?
View attachment 94777
Suggest the below...
Binned implementation, or data binning, is a pre-processing technique used in data science, machine learning, and simulations to group continuous, numerical data into discrete intervals (bins). This approach reduces the effects of minor observation errors and simplifies data analysis by replacing raw data points with their corresponding bin indices or averages.
Common Binned Implementations & Use Cases
- Machine Learning (Preprocessing): Discretizing numerical features using methods like equal-width (dividing range into equal sizes) or frequency binning (assigning similar number of values to each bin) to handle outliers or improve model robustness.
- Self-Supervised Learning (SSL): Reconstructing bin indices rather than raw values as a pretext task to enhance representation learning in tabular domains.
- Statistical Analysis: Computing binned Kernel Density Estimation (KDE) to estimate proper motions in astronomy.
- Sensitivity Analysis (Dakota): Using binned approaches to compute Sobol' main effect indices without needing a specific sampling structure.
- Physical Modeling: Implementing binned collision-coalescence models in meteorology to better simulate cloud droplet growth compared to bulk methods.
But they can't bury you if your heart's still beating ...
Is this a clip from a movie?View attachment 94778
So now it's becoming a reality soon
Pity its limited to software efficiency only.This is going to be cool to watch play out.
https://theaigrandprix.com
Andurill Industries
Today we’re announcing the AI Grand Prix. The fully autonomous drone racing competition inviting the boldest engineers from around the globe to compete for $500,000 and a job at Anduril.
No human pilots.
No hardware mods.
Identical Neros Tech drones.
Software is the only path to victory.
If you win, it’s because your autonomy stack is better. Full stop.
Season 1 kicks off this spring, leading up to the AI Grand Prix Ohio.
There is no link for the competition to BrainChip And the drones will be built by Neros Technologies.
“Neros Technologies (neros.tech) is a relatively new U.S.-based defense tech startup founded in 2023, specializing in manufacturing high-performance, attritable unmanned aerial systems (drones/FPV drones) for military use. It's focused on onshoring production, NDAA-compliant systems, electronic warfare resilience (e.g., fiber-optic variants like Archer Fiber), and battlefield-tested tech (including deployments in Ukraine). ”
Perhaps a player in the competition will use BrainChip software?
I believe that is from the second movie of The Kingsmen.View attachment 94778
So now it's becoming a reality soon