1
Kernels Benchmarks
⏲
Manage and navigate through HTML content within an iframe
ABDENNACER BADAOUI
badaoui
AI & ML interests
None yet
Recent Activity
new activity
2 minutes ago
badaoui/Tcid:fix data collection
liked
a Space
about 4 hours ago
kernels-community/kernels-benchmarks
updated
a Space
about 18 hours ago
badaoui/Tcid
Organizations
fix data collection
#1 opened 2 minutes ago
by
badaoui
liked
a
Space
about 4 hours ago
updated
a
Space
about 18 hours ago
updated
a
Space
6 days ago
Add link to Neuron-optimized version
#14 opened 7 days ago
by
badaoui
updated
a
model
7 days ago
published
a
model
7 days ago
feat: add progress indicator while compiling is running
1
#2 opened 7 days ago
by
tengomucho
reacted to
abidlabs's
post with 👍
8 days ago
Post
7405
Why I think local, open-source models will eventually win.
The most useful AI applications are moving toward multi-turn agentic behavior: systems that take hundreds or even thousands of iterative steps to complete a task, e.g. Claude Code, computer-control agents that click, type, and test repeatedly.
In these cases, the power of the model is not how smart it is per token, but in how quickly it can interact with its environment and tools across many steps. In that regime, model quality becomes secondary to latency.
An open-source model that can call tools quickly, check that the right thing was clicked, or verify that a code change actually passes tests can easily outperform a slightly “smarter” closed model that has to make remote API calls for every move.
Eventually, the balance tips: it becomes impractical for an agent to rely on remote inference for every micro-action. Just as no one would tolerate a keyboard that required a network request per keystroke, users won’t accept agent workflows bottlenecked by latency. All devices will ship with local, open-source models that are “good enough” and the expectation will shift toward everything running locally. It’ll happen sooner than most people think.
The most useful AI applications are moving toward multi-turn agentic behavior: systems that take hundreds or even thousands of iterative steps to complete a task, e.g. Claude Code, computer-control agents that click, type, and test repeatedly.
In these cases, the power of the model is not how smart it is per token, but in how quickly it can interact with its environment and tools across many steps. In that regime, model quality becomes secondary to latency.
An open-source model that can call tools quickly, check that the right thing was clicked, or verify that a code change actually passes tests can easily outperform a slightly “smarter” closed model that has to make remote API calls for every move.
Eventually, the balance tips: it becomes impractical for an agent to rely on remote inference for every micro-action. Just as no one would tolerate a keyboard that required a network request per keystroke, users won’t accept agent workflows bottlenecked by latency. All devices will ship with local, open-source models that are “good enough” and the expectation will shift toward everything running locally. It’ll happen sooner than most people think.
- 7 replies
upvoted
an
article
8 days ago
Article
Get your VLM running in 3 simple steps on Intel CPUs
•
18
Add link to Neuron-optimized version
#14 opened 9 days ago
by
badaoui
updated
a
model
9 days ago
published
a
model
9 days ago
liked
a
Space
9 days ago
Add historical data visualization features
🚀
🔥
1
3
#7 opened 13 days ago
by
badaoui
liked
a
Space
13 days ago
Add historical data visualization features
🔥
🚀
1
3
#7 opened 13 days ago
by
badaoui
updated
2
Spaces
14 days ago