Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
ArtificialLearn
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
4 ms
·
1.
▲
by
ArtificialLearn
12y ago
The Stanford chips are biomimetic, trying to copy how the brain is thought to work. Artificial Learning's are machine-learning theoretic, embodying proven ML algorithms in hardware.
2.
▲
by
ArtificialLearn
12y ago
Artificial Learning's largest designs are 180nm tech so some way from Moore's Law limits. Efficiency advantage over GPGPU implementation calculated as better than 10,000 fold in terms of learning updates per second per watt per k