Dustin Irwin

Data science, AI/ML tech, Julia language, Magic: the Gathering enthusiast

[Logan Kilpatrick @OfficialLoganK]

Gemini is officially coming to @GitHub Copilot!

Starting with Gemini 1.5 Pro, developers will get access to the 2 million token context window and native multimodal capabilities. It’s time to build.

https://t.co/e4OarEuypC

[Dustin Irwin @DustyIrwin]

@SaltyGoat17 @elonmusk Not sure I'd trust anyone who judges the value of ideas based solely on where they are discussed.

[Dustin Irwin @DustyIrwin]

@elonmuskADO

[Yann LeCun @ylecun]

🥁 Llama3 is out 🥁
8B and 70B models available today.
8k context length.
Trained with 15 trillion tokens on a custom-built 24k GPU cluster.
Great performance on various benchmarks, with Llam3-8B doing better than Llama2-70B in some cases.
More versions are coming over the next few months.

https://t.co/EkU9aIHdZE

[Logan Kilpatrick @OfficialLoganK]

Llama 3 is now available on @kaggle 🦙 congrats to the Meta team on such an impressive launch!

https://t.co/sX5OFLVqw9

[Logan Kilpatrick @OfficialLoganK]

We just dropped 2 new @Google Gemma open models.

Based on our original pre-trained Gemma models, CodeGemma brings powerful code completion and generation.

RecurrentGemma leverages RNNs and local attention to improve efficiency. Check it out!

https://t.co/Flx4DiRMTx

[Dustin Irwin @DustyIrwin]

@OfficialLoganK Have these companies to turn up the temperature on the AI models generating these pitch decks. ;)

[Logan Kilpatrick @OfficialLoganK]

Good news for @OpenAIDevs:

The Whisper API now support word and segment level timestamps which makes editing audio / video much easier and granular 🎬

https://t.co/xm2XYygFn5

[Dr. Chris Rackauckas @ChrisRackauckas]

New open source tool from @JuliaHub_Inc: static code analysis to prove that a #julialang code is allocation-free. Use this to ensure that codes are safe for real-time applications, such as how we use it for JuliaSim to analyze #SciML control codes!

https://t.co/9MpQzgLGqS

[Dustin Irwin @DustyIrwin]

@StefanKarpinski Whoa, that is terrible. ANY EU gov entity can basically decrypt your web traffic... This has me thinking what kind of security is in place for US traffic...

[Dustin Irwin @DustyIrwin]

@hus_qy Thanks for the update and your take on this 'organic'? approach to updating the iota-core logic. I look forward to reading more.

[Dustin Irwin @DustyIrwin]

@IBreakemoffI @Harada_TEKKEN Disagree. I enjoy interaction with my opponent much more than watching their helpless body flop around screen. IMO, it's boring for both sides.

[Dustin Irwin @DustyIrwin]

@muggleraph @algunion Couldn't agree more. With the code caching features in 1.9, I hope we're getting closer to easy to create standalone executables. I'm hopeful that the 1.10 release will focus on this issue

[Dustin Irwin @DustyIrwin]

@mtg_hotdog This is a nice resource sink that gives card advantage and doesn't tap--so it's abusable. Mmm delish!

[JOSΞPH 🌳 @CupOJoseph]

QR codes are about to get an upgrade. They work

[Dustin Irwin @DustyIrwin]

@elonmusk @davidmweissman @Target Oh sure, you're just out here asking the real questions.

[Dustin Irwin @DustyIrwin]

@ChrisRackauckas This is wonderful! I might even be able to disable my error_filter funcs :)

[Dustin Irwin @DustyIrwin]

For the folks I've recommended to try out Whisper lately, it looks like there is yet another AI library to consider for your project. 😅

[Yann LeCun @ylecun]

MMS: Massively Multilingual Speech.
- Can do speech2text and text speech in 1100 languages.
- Can recognize 4000 spoken languages.
- Code and models available under the CC-BY-NC 4.0 license.
- half the word error rate of Whisper.

Code+Models: https://t.co/NIGfUZ8KZg
Paper: https://t.co/W15aEWHGIR
Blog: https://t.co/TFKXFtlPwc

[Dustin Irwin @DustyIrwin]

@OfficialLoganK @OpenAI Thanks for what you do! Please don't forget to take time out for yourself as well.