Applied Intuition research develops LFG model to learn driving from internet dashcam footage, not lidar.

{
  "signal_id": "29aa6cb7-b746-4ccb-b521-1432cb852483",
  "signal_type": "linkedin-post-company",
  "signal_subtype": "linkedinPost",
  "detected_at": "2026-08-13T15:49:03.256+00:00",
  "company": {
    "name": "Applied Intuition",
    "domain": "appliedintuition.com"
  },
  "data": {
    "tags": [
      "Artificial Intelligence",
      "Research & Development",
      "Autonomous Technology",
      "Data Management",
      "Software"
    ],
    "summary": "Applied Intuition research develops LFG model to learn driving from internet dashcam footage, not lidar.",
    "post_url": "https://www.linkedin.com/posts/applied-intuition-inc_most-autonomy-pretraining-still-depends-on-activity-7492610015577448448-Epqu",
    "num_likes": 165,
    "post_text": "Most autonomy pretraining still depends on data like lidar scans, HD maps, and hand-annotated trajectories and labels. Our research team developed LFG (Learning to Drive is a Free Gift) to question that norm: what if a model could learn to drive by watching the internet's dashcam footage instead?\n\nThe results:\n\n🔶 Trained on ~2M unlabeled dashcam clips using a teacher-student setup, no ground-truth labels required\n🔶 Surpasses its own teacher on future-frame semantic segmentation, despite predicting frames it never saw\n🔶 Hits 85.2 PDMS on NAVSIM with a single front-facing camera, outperforming multi-camera and lidar-equipped systems like UniAD and Hydra-MDP\n\nBy learning from freely available internet footage, LFG sidesteps the data bottleneck that has constrained previous approaches.\n\nRead the full paper breakdown: https://lnkd.in/g8_ywXtx",
    "initiatives": [
      {
        "topic": "developing LFG model for autonomous driving",
        "urgency": 0.8
      },
      {
        "topic": "researching driving models from internet footage",
        "urgency": 0.7
      }
    ],
    "pain_points": [
      {
        "topic": "autonomy pretraining data bottleneck",
        "intensity": 0.7
      }
    ],
    "posted_date": "2026-08-10T15:57:13.676Z",
    "num_comments": 5,
    "competitors_mentioned": [
      {
        "name": "UniAD"
      },
      {
        "name": "Hydra-MDP"
      }
    ],
    "technologies_mentioned": [
      {
        "name": "lidar",
        "status": "evaluating"
      },
      {
        "name": "semantic segmentation",
        "status": "using"
      }
    ]
  }
}

curl https://signals.autobound.ai/v1/signals/29aa6cb7-b746-4ccb-b521-1432cb852483 \
  -H "X-API-KEY: $AUTOBOUND_API_KEY"

curl -X POST https://signals.autobound.ai/v1/companies/enrich \
  -H "X-API-KEY: $AUTOBOUND_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"domain":"appliedintuition.com","limit":20}'