# \#distillation

**URL:** https://community.deeplearning.ai/tag/distillation/118.md

[Latest](https://community.deeplearning.ai/latest.md) · [Categories](https://community.deeplearning.ai/categories.md) · [Tags](https://community.deeplearning.ai/tags.md)

---

## [I Built a New Software Primitive in 8.5 Hours. It Replaces the Eyes of Every AI Agent on Earth](https://community.deeplearning.ai/t/i-built-a-new-software-primitive-in-8-5-hours-it-replaces-the-eyes-of-every-ai-agent-on-earth/889040)

<div class="topic-metadata">

**Author:** [@Martin\_Gehrken](https://community.deeplearning.ai/u/Martin_Gehrken)\
**Replies:** 1\
**Last updated:** [April 7, 2026, 7:59pm UTC](https://community.deeplearning.ai/t/i-built-a-new-software-primitive-in-8-5-hours-it-replaces-the-eyes-of-every-ai-agent-on-earth/889040 "2026-04-07T19:59:45Z")

</div>

DirectShell: I Turned the Accessibility Layer Into a Universal App Interface. No Screenshots. No Vision Models. Martin Gehrken — February 17, 2026 As of February 17, 2026, every screenshot-based AI agent, every enterpr…

---

## [Week 3: Distillation: Train a big model, then use that the train a small model, seems convoluted?](https://community.deeplearning.ai/t/week-3-distillation-train-a-big-model-then-use-that-the-train-a-small-model-seems-convoluted/706671)

<div class="topic-metadata">

**Author:** [@Peppers16](https://community.deeplearning.ai/u/Peppers16)\
**Replies:** 3\
**Last updated:** [January 25, 2025, 10:28pm UTC](https://community.deeplearning.ai/t/week-3-distillation-train-a-big-model-then-use-that-the-train-a-small-model-seems-convoluted/706671 "2025-01-25T22:28:10Z")

</div>

In Week 3 when learning about deploying performant models, one method that was discussed was “distillation”. In order to get a smaller, less expensive model: we take our trained, fine-tuned, RLHF-trained model… then we …