---
title: "Logit Solutions AI (@logitsolutionsai) | Hashnode"
description: "AI Products That Work in the Real World. Read the latest articles by Logit Solutions AI on Hashnode."
url: "https://hashnode.com/@logitsolutionsai"
---

@logitsolutionsai

# Logit Solutions AI

@logitsolutionsai · Joined October 2026

AI Products That Work in the Real World.

[Website](https://logitsolutions.ai/)

### [Speech-to-text can't tell you if an AI voice said a name right](https://logitsolutionsai.hashnode.dev/speech-to-text-can-t-tell-you-if-an-ai-voice-said-a-name-right?utm_source=hashnode&utm_medium=feed)

We make short, AI-voiced lessons. Before a video ships, every sentence goes through automatic checks, and then a person listens. The automatic checks kept failing in one place: names. This post explai

5 min read · [logitsolutionsai.hashnode.dev](https://logitsolutionsai.hashnode.dev/speech-to-text-can-t-tell-you-if-an-ai-voice-said-a-name-right?utm_source=hashnode&utm_medium=feed)

[3](https://hashnode.com/posts/speech-to-text-can-t-tell-you-if-an-ai-voice-said-a-name-right/6af3a86b8e08d0169e437c33) [10h ago](https://hashnode.com/posts/speech-to-text-can-t-tell-you-if-an-ai-voice-said-a-name-right/6af3a86b8e08d0169e437c33)

### [Batch normalization, visually: the reset that keeps deep networks steady](https://logitsolutionsai.hashnode.dev/batch-normalization-visually-the-reset-that-keeps-deep-networks-steady?utm_source=hashnode&utm_medium=feed)

Deep Learning, Visually #20 Watch the Short: youtube.com/shorts/yYxpjlvW69s Why the numbers drift A deep network is a chain of layers. Each layer takes a weighted sum of the activations from the layer

5 min read · [logitsolutionsai.hashnode.dev](https://logitsolutionsai.hashnode.dev/batch-normalization-visually-the-reset-that-keeps-deep-networks-steady?utm_source=hashnode&utm_medium=feed)

[1d ago](https://hashnode.com/posts/batch-normalization-visually-the-reset-that-keeps-deep-networks-steady/6af200bd3bb56a980ff1851a)

### [Why did deep networks stop learning?](https://logitsolutionsai.hashnode.dev/why-did-deep-networks-stop-learning?utm_source=hashnode&utm_medium=feed)

Vanishing and exploding gradients: why blame fades or blows up on its way back through a deep network, and the fixes that made deep learning work

5 min read · [logitsolutionsai.hashnode.dev](https://logitsolutionsai.hashnode.dev/why-did-deep-networks-stop-learning?utm_source=hashnode&utm_medium=feed)

[6d ago](https://hashnode.com/posts/why-did-deep-networks-stop-learning/6aeb40777750cbafe81b0f82)

### [Why does every AI start out random?](https://logitsolutionsai.hashnode.dev/why-does-every-ai-start-out-random?utm_source=hashnode&utm_medium=feed)

Weight initialization: why a network's dials must start different, and how big the random start should be

6 min read · [logitsolutionsai.hashnode.dev](https://logitsolutionsai.hashnode.dev/why-does-every-ai-start-out-random?utm_source=hashnode&utm_medium=feed)

[6d ago](https://hashnode.com/posts/why-does-every-ai-start-out-random/6aeb4039e748939f844415fe)

### [Why slow down near the finish line?](https://logitsolutionsai.hashnode.dev/why-slow-down-near-the-finish-line?utm_source=hashnode&utm_medium=feed)

Learning-rate schedules: start big, halve the step size every so often, and settle where one step size can't

6 min read · [logitsolutionsai.hashnode.dev](https://logitsolutionsai.hashnode.dev/why-slow-down-near-the-finish-line?utm_source=hashnode&utm_medium=feed)

[Nov 1](https://hashnode.com/posts/why-slow-down-near-the-finish-line/6ae74da0f2fb45ad8f13e385)