SEQUOIA: Exact Llama2-70B on an RTX4090 with half-second per-token latency
Serving Llama2-70B on an RTX4090 with Sequoia
The latest blog posts to take off on Hacker News. 1,840 posts.
Serving Llama2-70B on an RTX4090 with Sequoia
How we built a performant, scalable, ACID-compliant, JavaScript-native database on FoundationDB.
My goal with this bookmarkable guide is to provide a list of (frankly: incredible) new additions to CSS lately. There is no hardline criteria for this list other than that these things are all fairly new and my sense is that many people aren't aware of these things. Or even if they are, they don't…
We were recently made aware of multiple potential DNS leaks on Android. They stem from bugs in Android itself, and only affect certain apps.
Creating a low-cost EV charging station with Arduino: see how a maker built a smart charger for electric vehicles.
Figma's team recently converted one of its codebases from a custom programming language to TypeScript without disrupting a single day of development.
From API contract to documentation portal: streamline your workflow and deliver the best API experience. Built for engineers and tech writers.
To avoid catastrophic climate impacts, excessive carbon emissions must be addressed. At this point, cutting emissions isn’t enough. Direct air capture, a technology that pulls carbon dioxide out of ambient air, has great potential to help solve the problem.
Emerge Tools analysis of the Rabbit R1 APK.
How I learned to stop worrying and love the burn.
Why gene banks aren’t enough to save the world’s food
Analyze Backblaze's Q1 2024 hard drive stats, including annualized failure rates, average ages of failure, and insights on over 279,000 drives.
Information on a security incident involving Dropbox Sign.
If you publicly host large data files on AWS S3 & pay for transfer costs, you may be vulnerable to a “Denial of Wallet” amplification…
Go 1.22 adds math/rand/v2 and charts a course for the evolution of the Go standard library.
The EU has outperformed the US on per-capita output growth; in terms of output per hour worked, some EU countries are as productive as the US
Determination must be reviewed by the White House Office of Management and Budget
Read about the challenges, pitfalls and misconceptions we faced while developing our passkey-first authentication solution and passkey intelligence.
It took more than ten years of full-time work for Rama to go from an idea to a production system. I shudder to think of how long it would have taken without Clojure. Rama is a programming platform …
Researchers at DTU and Lund University have discovered enzymes that remove hindrances that stood in the way of developing universal donor blood.
OpenTofu 1.7.0 is now available with full state encryption, dynamic provider-defined functions, the removed and loopable import blocks, new migration guides, and much more.
Every SQL database engine works in roughly the same way: It first translates the input SQL text into a "prepared statement". Then it "executes" the prepared statement to generate a result.
We’re redefining the developer environment with GitHub Copilot Workspace - where any developer can go from idea, to code, to software all in natural language.
The NASA/ESA/CSA James Webb Space Telescope has captured the sharpest infrared images to date of one of the most distinctive objects in our skies, the Horsehead Nebula. These observations show a part of the iconic nebula in a whole new light, capturing its complexity with unprecedented spatial resolution.
By imposing the restrictions on open-source AI, SB-1047 could reduce AI safety, through reducing transparency, collaboration, diversity, and resilience.
How to infer character encodings with statistics
The "thorium transition", which physicists have been looking for for decades, has now been excited for the first time with lasers. This paves the way for revolutionary high precision technologies, including nuclear clocks.
A USPS smishing attempt prompts the Akamai SIG to unearth shocking statistics: Illegitimate USPS domains see as much traffic as the legitimate USPS domain.