Cortensor
@cortensor
Followers
4K
Following
15K
Media
423
Statuses
1K
Pioneering Decentralized AI Inference, Democratizing Universal Access. #AI #DePIN 🔗 https://t.co/eVDxab7WYU
Joined June 2024
🗓️ Weekly Recap – L3, SLA, and Phase #5 Prep. 🔹 Cognitive Level / SLA-Based Node Assignment Functional.- NodePool now integrates SLA flags to select nodes by real-time capability. - Full range from L0–L10 supported with dynamic tier detection. - Further optimization & adaptive
🗓️ Weekly Focus – Wrapping Up & Gearing Forward. 🔹 Finalizing Phase #4 prize, July node rewards & hackathon payouts.- Including EarlyBird4 role assignment and reward tier processing across all eligible nodes. 🔹 Tuning DevNet6 and prepping for Phase #5 launch.- Live debugging
6
10
25
Monthly Node Rewards – July 2025 (Batch 6 Sent). Another 23 active nodes have just received their $COR rewards - bringing the total to 173 rewarded nodes so far this month. 🔹Rewards will be distributed in batches.🔹More batches rolling out over the next few days.🔹Recognizing
Monthly Node Rewards – July 2025 (Batch 5 Sent). Another 30 active nodes have just received their $COR rewards - bringing the total to 150 rewarded nodes so far this month. 🔹Rewards will be distributed in batches.🔹More batches rolling out over the next few days.🔹Recognizing
2
4
15
🛠️ DevLog – Smarter SLA Node Selection (v1). Further testing led to key improvements in how nodes are matched to SLA tiers. This update addresses an important issue where top-tier nodes (e.g., L6) weren't being properly assigned. 🔹 Dynamic Range Detection.- Detects actual.
🛠️ DevLog – SLA Node Selection Fix Deployed. 🔹 Issue Identified.- High SLA sessions were sometimes matched with low-level nodes.- Target level exceeded available nodes (e.g., requested level 8, highest was 6).- System fell back to random low-level nodes instead of best.
4
8
12
🛠️ DevLog – SLA Node Selection Fix Deployed. 🔹 Issue Identified.- High SLA sessions were sometimes matched with low-level nodes.- Target level exceeded available nodes (e.g., requested level 8, highest was 6).- System fell back to random low-level nodes instead of best.
🛠️ DevLog – SLA-Based Node Selection via Cognitive Level Integration. Cortensor's Node Pool now supports SLA-aware node selection using Cognitive Level (L0–L6) categories. 🎯. With this update:.- Sessions can now specify an SLA Level (Low, Medium, High). - During task submission,
3
10
22
🛠️ DevLog – Dashboard Refactor for Dual Network Support. As part of the L3 AppChain rollout (where $COR is used as gas for node operations), we're refactoring the dashboard to support both:. 🔹 Single-network mode (Testnet #0 – Arbitrum Sepolia only).🔹 Dual-network mode (Testnet.
🛠️ DevLog – L3 ($COR as Gas) Ready for Stress Testing. Cortensor's full stack is now deployed on L3 Rollup with $COR as native gas - and we're gearing up for live stress tests. 🔹 L2 → L3 Bridge Working.Bridge from Arbitrum Sepolia to L3 is now functional, and token transfers
4
11
22
Monthly Node Rewards – July 2025 (Batch 5 Sent). Another 30 active nodes have just received their $COR rewards - bringing the total to 150 rewarded nodes so far this month. 🔹Rewards will be distributed in batches.🔹More batches rolling out over the next few days.🔹Recognizing
Monthly Node Rewards – July 2025 (Batch 4 Sent). Another 30 active nodes have just received their $COR rewards - bringing the total to 120 rewarded nodes so far this month. 🔹Rewards will be distributed in batches.🔹More batches rolling out over the next few days.🔹Recognizing
0
2
17
🛠️ DevLog – SLA-Based Node Selection via Cognitive Level Integration. Cortensor's Node Pool now supports SLA-aware node selection using Cognitive Level (L0–L6) categories. 🎯. With this update:.- Sessions can now specify an SLA Level (Low, Medium, High). - During task submission,
🛠️ DevLog – L3, NodeLevel, and SLA Integration. L3 Rollup w/ $COR as gas continues to run smoothly.→ → 2 Oracles + 2 Miners live to test full-stack flow. 🔹 Major focus this week: L3 Rollup w/ $COR as gas & integrating CognitiveLevel / NodeLevel and.
3
8
16
Monthly Node Rewards – July 2025 (Batch 4 Sent). Another 30 active nodes have just received their $COR rewards - bringing the total to 120 rewarded nodes so far this month. 🔹Rewards will be distributed in batches.🔹More batches rolling out over the next few days.🔹Recognizing
Monthly Node Rewards – July 2025 (Batch 3 Sent). Another 30 active nodes have just received their $COR rewards - bringing the total to 90 rewarded nodes so far this month. 🔹Rewards will be distributed in batches.🔹More batches rolling out over the next few days.🔹Recognizing
2
6
17
🛠️ DevLog – L3, NodeLevel, and SLA Integration. L3 Rollup w/ $COR as gas continues to run smoothly.→ → 2 Oracles + 2 Miners live to test full-stack flow. 🔹 Major focus this week: L3 Rollup w/ $COR as gas & integrating CognitiveLevel / NodeLevel and.
testnet1.explorer.cortensor.network
🗓️ Weekly Focus – Wrapping Up & Gearing Forward. 🔹 Finalizing Phase #4 prize, July node rewards & hackathon payouts.- Including EarlyBird4 role assignment and reward tier processing across all eligible nodes. 🔹 Tuning DevNet6 and prepping for Phase #5 launch.- Live debugging
4
7
13
🛠️ DevLog – Expanding Model Support for User Task Mode. Two new models are now supported for dedicated node setups in User Task Mode, available in both subprocess and Docker modes:. 🔹 DeepSeek-R1-Distill-Llama-8B-Q4_K_M.🔹 Meta-Llama-3.1-8B-Instruct.Q4_K_M. Even though these are.
🗓️ Weekly Focus – Wrapping Up & Gearing Forward. 🔹 Finalizing Phase #4 prize, July node rewards & hackathon payouts.- Including EarlyBird4 role assignment and reward tier processing across all eligible nodes. 🔹 Tuning DevNet6 and prepping for Phase #5 launch.- Live debugging
4
9
17
Monthly Node Rewards – July 2025 (Batch 3 Sent). Another 30 active nodes have just received their $COR rewards - bringing the total to 90 rewarded nodes so far this month. 🔹Rewards will be distributed in batches.🔹More batches rolling out over the next few days.🔹Recognizing
Monthly Node Rewards – July 2025 (2nd Batch Sent). We've just sent out the 2nd round of $COR rewards to 30 active nodes, as part of our monthly node reward program. 🔹 Rewards will be distributed in batches.🔹 More batches rolling out over the next few days.🔹 Recognizing
6
11
21
🛠️ DevLog – SLA-Aware Sessions Live, Node Routing Testing Next. We've redeployed the latest Session and NodePool modules on DevNet6, introducing SLA flags (low, medium, high) into session creation - aligned with our CognitiveLevel (Node Level) architecture. 🔹 Session SLA Flag
🛠️ DevLog – SLA-Based Scheduling (WIP): Session + NodePool Refactor Underway. We're preparing a major update to enable SLA-aware task scheduling powered by CognitiveLevel. This will introduce more intelligent and performance-matched routing. 🔹 New Flow Overview:.1. User creates.
4
7
14
🛠️ DevLog – L3 ($COR as Gas) Ready for Stress Testing. Cortensor's full stack is now deployed on L3 Rollup with $COR as native gas - and we're gearing up for live stress tests. 🔹 L2 → L3 Bridge Working.Bridge from Arbitrum Sepolia to L3 is now functional, and token transfers
🛠️ DevLog – L3 with $COR as Gas Now Live (Experimental). We've deployed the experimental L3 rollup with native $COR as gas – part of our long-term appchain alignment. 🔹 Network: cortensor-testnet-1.🔹 Explorer: 🔹 COR bridged from L1 → L2 → L3.🔹.
2
5
16
🛠️ DevLog – SLA-Based Scheduling (WIP): Session + NodePool Refactor Underway. We're preparing a major update to enable SLA-aware task scheduling powered by CognitiveLevel. This will introduce more intelligent and performance-matched routing. 🔹 New Flow Overview:.1. User creates.
🛠️ DevLog – CognitiveLevel Iteration: Capped Levels, SLA Categories, and Node-Aware Scheduling. As part of our ongoing CognitiveLevel iteration:. 🔹 Level Cap Introduced:.We've capped levels to 0–5. Submitting recursive inference beyond Level 5 exceeds onchain tx size limits - so
3
11
25
🛠️ DevLog – CognitiveLevel Iteration: Capped Levels, SLA Categories, and Node-Aware Scheduling. As part of our ongoing CognitiveLevel iteration:. 🔹 Level Cap Introduced:.We've capped levels to 0–5. Submitting recursive inference beyond Level 5 exceeds onchain tx size limits - so
🛠️ DevLog – Recursive Inference Enabled on Miner Runtime for Node Level. Upgrade & downgrade logic is now live on DevNet6 – triggered after 10+ successful tasks with 90%+ accuracy. 🔹 Miner runtime now pulls CognitiveLevel.🔹 Based on assigned level, miners run recursive
3
6
22
Monthly Node Rewards – July 2025 (2nd Batch Sent). We've just sent out the 2nd round of $COR rewards to 30 active nodes, as part of our monthly node reward program. 🔹 Rewards will be distributed in batches.🔹 More batches rolling out over the next few days.🔹 Recognizing
Monthly Node Rewards – July 2025 (Initial Batch Sent). We've just sent out the first round of $COR rewards to 30 active nodes, as part of our monthly node reward program. 🔹 Rewards will be distributed in batches.🔹 More batches rolling out over the next few days.🔹 Recognizing
1
3
13
🛠️ DevLog – L3 with $COR as Gas Now Live (Experimental). We've deployed the experimental L3 rollup with native $COR as gas – part of our long-term appchain alignment. 🔹 Network: cortensor-testnet-1.🔹 Explorer: 🔹 COR bridged from L1 → L2 → L3.🔹.
testnet1.explorer.cortensor.network
4
9
21
🛠️ DevLog – Recursive Inference Enabled on Miner Runtime for Node Level. Upgrade & downgrade logic is now live on DevNet6 – triggered after 10+ successful tasks with 90%+ accuracy. 🔹 Miner runtime now pulls CognitiveLevel.🔹 Based on assigned level, miners run recursive
🛠️ DevLog – Validator v1 Level Adjustment Logic Live. We've kicked off upgrade/downgrade logic for Node Levels in DevNet6 via validator_v1. 🔹 How It Works:.- Validator now tracks each node’s task result history. - After 50 successful network tasks, upgrade/downgrade triggers.
3
9
27