# Mar 30, 2024

## Playboy image from 1972 gets ban from IEEE computer journals
- The **IEEE Computer Society** announced a ban on papers including the **"Lenna" image**, a cropped photo of a 1972 Playboy model, starting April 1, in an effort to promote a more inclusive and equitable culture.

## Qwen1.5-Moe: Matching 7B Model Performance with 1/3 Activated Parameters
- **Qwen1.5-MoE-A2.7B** matches the performance of **7B models** with **only 2.7 billion activated parameters**, achieving a **75% reduction in training costs** and a **1.74x speed increase in inference**.

## Towards 1-bit Machine Learning Models
- **Recent advancements in machine learning, such as [BitNet](https://arxiv.org/abs/2310.11453) and [1.58 bit](https://arxiv.org/abs/2402.17764), focus on extreme low-bit quantization**, aiming to implement matrix multiplication with quantized weights without multiplications, potentially revolutionizing compute efficiency.

## Navigating the Challenges and Opportunities of Synthetic Voices
- OpenAI's **Voice Engine** model generates **natural-sounding speech** from text and a 15-second audio sample, closely mimicking the original speaker's voice, with applications ranging from educational aids to therapeutic support.

## TnT-LLM: Text Mining at Scale with Large Language Models
- **TnT-LLM** leverages **Large Language Models (LLMs)** to **automate label generation and assignment**, streamlining the text mining process by reducing the need for domain expertise and manual curation.

## How Stability AI’s Founder Tanked His Billion-Dollar Startup
- **Stability AI's founder, Emad Mostaque, resigned** amid financial turmoil and internal pressure, following a series of poor business decisions and extravagant spending that eroded investor confidence and jeopardized the company's future.

## Jamba: the first production-grade Mamba-based model delivering best-in-class quality and performance.
- **Jamba** integrates **Mamba's Structured State Space model** with traditional Transformer architecture, enhancing efficiency and throughput while maintaining a **256K context window**.

## Launch HN: Lumona (YC W24) – Product search based on Reddit and YouTube reviews
- **Lumona** introduces a **product search engine** that leverages social media insights, specifically from **Reddit and YouTube**, to recommend skincare products, with plans for future expansion.

## The Tech Behind The Magic: How OpenAI SORA Works
- **OpenAI SORA** represents a significant leap in **image and video generation AI technologies**, showcasing advanced training methodologies and societal impacts.

## AutoBNN: Probabilistic Time Series Forecasting
- **AutoBNN merges the interpretability of probabilistic models with neural networks' scalability**, automating the creation of sophisticated time series forecasting models that are both interpretable and capable of handling large datasets efficiently.

## Nine writers on their Adderall use and experiences
- **America's Adderall usage** has surged, embedding a culture of prescription-grade speed among professionals and students, with the drug's popularity obscuring its addictive nature and health risks.

## How would you answer this interview question?
- **Predicting the probability** of the second-place car overtaking the first in an F1 race involves **analyzing historical data, current race dynamics, and driver performance**.

## Question about DDIM paper
- A **question** was raised about a potential error in the **DDIM paper**, specifically regarding the direction of an inequality that resembles the standard **ELBO**.

## Rustic UI: Crafting the Future of UX
- **Rustic UI** is a design framework aimed at enhancing **AI-driven conversational UX** through a multimodal approach, offering tools for personalized and accessible user experiences.

## Looking to build a personalized AI for myself. Any resources?
- The individual seeks to **develop a personalized AI** to enhance communication within an online community by understanding its members and suggesting conversation topics.