Julia (@juliafterwork) header
Julia (@juliafterwork) profile picture

Born after the World Wide Web. ⚠️ I might talk a lot about machine learning.

Earth
Joined March 2018
Photo by Julia (@juliafterwork)Photo by Julia (@juliafterwork)Photo by Julia (@juliafterwork)Photo by Julia (@juliafterwork)Photo by Julia (@juliafterwork)Photo by Julia (@juliafterwork)
Hello @langchain, We have 80 face-to-face participants for this meetup (sold out), and 420 remote participants. That's 500 people in all. -> This prompts the question: Could this be the largest meetup globally focused on your technology? <- Your technology is top-notch, it significantly advancing the LLM/GenAI ecosystem, bringing it to new levels. Thank you for your excellent work. meetup.com/fr-FR/langchain-a…
4
16
1,594
NEW: This is the device that confused and concerned attendees at @defcon by triggering pop ups on their iPhones. The researcher who created it explained how it works, and what are the risks iPhone users face when these spoofed messages pop up. techcrunch.com/2023/08/16/th…
12
284
26
835
161,677
Good work from a German ad agency which reads "protect the wildlife before it's too late"
102
10,361
372
37,989
2,126,139
No More GIL! the Python team has officially accepted the proposal. Congrats @colesbury on his multi-year brilliant effort to remove the GIL, and a heartfelt thanks to the Python Steering Council and Core team for a thoughtful plan to make this a reality. discuss.python.org/t/a-steer…
59
1,039
327
4,449
1,391,821
Most misconceptions about AI would be cleared if folks had a decent understanding of the distinction between skill and intelligence, and between memorization and generalization.
30
87
4
609
72,967
Excited to announce our multi-modal agent API that can automatically chain 100,000+ of HF models (stable diffusion, whisper, OpenAssistant,...) for text, audio, image, video, time-series,… based on your commands. All open-source so it can work locally! huggingface.co/docs/transfor…
25
162
23
708
105,278
La recherche sur Google "Regenerate response" site:.fr C'est fantastique #ChatGPT
19
178
29
815
356,762
The most powerful debugging tool: walking away from your computer.
67
330
54
2,522
187,117
Sometimes it feels like NLP papers prior to 2020 don't exist... (Bidirectional autoregressive models have been common for many years, and were for instance used in ULMFiT.)
4
5
2
148
42,203
Me, reading *peer-reviewed* papers that anthropomorphize AI models and say that they "choose" their own parameters.
15
28
4
317
36,118
If you have ideas of papers, authors, or topics you would like to see in this serie of #FridayWiMLDSPaper, send us a message!
2
2
166
It's our week #200 of #FridayWiMLDSPaper!! We are very proud of this accomplishment, and we would like to thank everyone who participated into it, among them: @bettymoreschini, @MarinaVinyes, @lrnt_chloe, @juliafterwork, @VictoireLouisC, @math_rachel, @ledell, and the team!
2
4
6
294
The tech industry prides itself on being data-driven, but for the most part it is moved by narratives, and those narratives are more often than not untethered from actual data.
41
130
28
963
110,318
Meta is really publish for Open Source ♥️ Actions speak louder than having "Open" in your company name
Announced by Mark Zuckerberg this morning — today we're releasing DINOv2, the first method for training computer vision models that uses self-supervised learning to achieve results matching or exceeding industry standards. More on this new work ➡️ https://nitter.cf/t.co/h5exzLJsFt
4
9
109
14,779
Thank you @ClementDelangue and @huggingface for #WoodstockAI! Haven’t seen this level of energy The LLaMas and walking emoji were just *chef’s kiss* #WoodstockAI #huggingface #llm #oss #ai #ml
8
2
95
170,218
As a specialist in evaluating language models, I declare that this is the best way of evaluating language models:
32
256
42
2,599
361,067
When I was a student, my performance would also increase significantly when I was able to procure the exam questions and their answers the week before.
I suspect GPT-4's performance is influenced by data contamination, at least on Codeforces. Of the easiest problems on Codeforces, it solved 10/10 pre-2021 problems and 0/10 recent problems. This strongly points to contamination. 1/4
16
136
9
1,519
238,764
All on one day: - OpenAI GPT-4 - Anthropic Claude - Google PaLM API & MakerSuite - Google PaLM-Med - Adept $350m round Huge day for AI.
2
5
35
3,141