Hacker News Top Stories with Summaries (December 17, 2023)
<style>
p {
font-size: 16px;
line-height: 1.6;
margin: 0;
padding: 10px;
}
h1 {
font-size: 24px;
font-weight: bold;
margin-top: 10px;
margin-bottom: 20px;
}
h2 {
font-size: 18px;
font-weight: bold;
margin-top: 10px;
margin-bottom: 5px;
}
ul {
padding-left: 20px;
}
li {
margin-bottom: 10px;
}
.summary {
margin-left: 20px;
margin-bottom: 20px;
}
</style>
<h1> Hacker News Top Stories</h1>
<p>Here are the top stories from Hacker News with summaries for December 17, 2023 :</p>
<div style="margin-bottom: 20px;">
<table cellpadding="0" cellspacing="0" border="0">
<tr>
<td style="padding-right: 10px;">
<div style="width: 200px; height: 100px; border-radius: 10px; overflow: hidden; background-image: url('https://spectrum.ieee.org/media-library/a-mottled-grey-slab-with-patterns-of-darker-and-lighter-grey-at-the-faces.jpg?id=50810401&width=1200&height=600&coordinates=0%2C137%2C0%2C138'); background-size: cover; background-position: center;">
Intel, Samsung, and TSMC Demo 3D-Stacked Transistors
Summary: Intel, Samsung, and TSMC have demonstrated complementary field-effect transistors (CFETs) at the IEEE International Electron Devices Meeting. CFETs are a single structure that stacks both types of transistors needed for CMOS logic, potentially doubling transistor density. Chipmakers are transitioning from FinFET to nanosheet or gate-all-around transistors, with CFETs expected to roll out commercially in seven to ten years.
<div style="margin-bottom: 20px;">
<table cellpadding="0" cellspacing="0" border="0">
<tr>
<td style="padding-right: 10px;">
<div style="width: 200px; height: 100px; border-radius: 10px; overflow: hidden; background-image: url('https://assets-global.website-files.com/6570a6bdf377183fb173431e/657358b490faaa82ad9138ca_open-graph.jpg'); background-size: cover; background-position: center;">
Transformers on Chips
Summary: Introducing Etched, the world's first transformer supercomputer with architecture etched into silicon chips for powerful transformer inference. It enables real-time voice agents, better coding with tree search, and multicast speculative decoding. The system supports trillion-parameter models, is fully open-source, and expandable to 100T param models. It features 144 GB HBM3E per chip and offers MoE and transformer variants. © 2023 Etched.