Yay - *free ebook* if you stay till the end! 👍👏🙏

great..

Love the slides and the explanation:)

Since many (all?) versions of Intel ME use MINIX , "MINIX *was* a small microkernel" is an understatement.

Amazing work! Dr. Nivedita ,keep it up ! stay Blessed, Madam for the ground breaking concern for mother Earth !👍

2024!

It's knowing what a woman is.

People need to stop being so scared of words. Normal is a perfectly acceptable way to describe a group of people and exclude a different group. Normal simply means you don’t require additional tools or services to get along in an environment. With regard to where I work, I’m not normal because I require additional approvals and paperwork to be allowed to have a CGM and insulin pump. A normal person, one that doesn’t require these special tools and devices, doesn’t need additional privileges to work where I work. Someone in a wheelchair would be abnormal where I work, because the infrastructure of the building would need to be drastically changed to accommodate their needs. Words are not scary and evil, it’s the way we communicate ideas. If someone is using words to attack you, then just ignore them. Who cares what that person thinks.

Salton Award lecture from Ellen Voorhees starts at 35'30''

One of the best sessions I ever had.

Thanks.

Thank you to all the volunteers who put this together :) I'm sure it was a lot of work

Catch an exclusive interview with Binance’s CEO to get the scoop on future developments

Due to some AV issues, I could not hear the questions clearly on the stage. So, let me clarify some of the answers retrospectively. Question 1. Why did the easy-to-heard voting strategies perform worse than the SFT baseline with greedy decoding when the candidate-pool size is less than 10? Answer: The voting strategies had a non-deterministic process, with the temperature set to 0.7 (as shown in the slide) while the baseline has the temperature set to 0 (deterministic). This means that the baseline always picked its top candidate per problem instance, but the voting strategies may miss it when the pool size is rather small. Question 2. How does the proposed method differ from GAN? Answer: The discriminator in GAN is trained to label each instance as a natural-language output (yes) or un-natural (no), while the evaluator in easy-to-hard generalization is trained to tell whether a math solution is correct (yes) or wrong (no) for a given math problem. Even if an answer is perfect in English, it still can be wrong mathematically. Question 3. Can we use this idea to improve LLM pre-training? Answer: Maybe not. When we have enough (unlabeled) data for pre-training of an LLM, we may not be benefited much from the easy-to-hard generalization. On the other hand, if we do not have enough data for pretraining an LLM, we may also not be able to train the evaluator well. One may argue that what if we can train the evaluator on rare patterns that the current LLM cannot handle well? Perhaps yes, but this is a big “if”. That is, the challenge is then shifted to how to obtain the annotated data on the rare patterns. I hope the above answers help.

How did I get here

Thanks.

8:59 Stream starts

Worst workflow. Very frustating

Congratulations! A well-deserved prize.

legend

real man !

great paper!

This new workflow template is a mess...

youtube limits download rate server side to about 1.2 times the bitrate of the video by setting a cwin flag on the TCP socket!

Asesome Prezzo❤, Is Moka an open source tool?

Unfortunately, this is a glimpse of the horror of The age.

Thanks for publicly sharing the talk; Avi is inspiring for all generations.

Introduction at 6:40 and talk starts at 10:45.

Amazing❤

Thanks. 😊

Thank you for share!

Yes I learn lots from the original Minix book all those years ago. Thank you Andrew !

28:51 Kostya’s talk: kzbin.info/www/bejne/p4XOind3ZaqhbLc

This is fascinating and I learned so much. I was only marginally aware of those branches of research and the developments, but it's so good to be able to listen to these wonderful people sharing their research and the important work they do. Absolutely fantastic and so very important.

Interesting panel. I love that research like this is being done. So important.

No to lgbt

Funny that Tulsee Doshi was on here talking about Google and fairness, only to have Gemini shown to have the most egregious bias ever seen in Artificial Intelligence.

There are unknown way to visualize subspace, or vector spaces. You can stretching the width of the x axis, for example, in the right line of a 3d stereo image, and also get depth, as shown below. L R |____| |______| TIP: To get the 3d depth, close one eye and focus on either left or right line, and then open it. This because the z axis uses x to get depth. Which means that you can get double depth to the image.... 4d depth??? :O It's even possible to add 4d depth to a isometric cube, by moving the skew square side along the x axis of the right cube, individually, and leave the left cube unchanged. Would be awesome, if someone make a game (even KZbin video) that way! Nice video! :-)

Very nice and profound Talk. Thanks Hannes and Thanks for ACM for sharing this publically on KZbin.

Keswick.

Wow! This is such a good introduction to the field. Thank you for this❤

Creo que se paso de ver... con lo de sasuke

🎯 Key points for quick navigation: Understanding anti-patterns in student code can indicate knowledge gaps and lack of understanding of correct patterns. Code writing alone may not provide a complete picture of a student's knowledge, as students can generate code without understanding underlying principles. Students who use anti-patterns in code writing may not necessarily have conceptual gaps, but may lack understanding of specific patterns. Code writing is weakly correlated with editing success, suggesting that additional tasks besides code writing may be necessary to assess student knowledge accurately. Engaging students emotionally with real-world examples, such as accidents caused by suboptimal software testing, can enhance learning and motivation. Learning from accidents in software engineering can help students understand the importance of safety and domain knowledge in software development. Encouraging witnesses of accidents to share their experiences with software engineering students can lead to better education and training of future software engineers. Students in the University course on mlops are taught how to adopt mlops practices and gain practical experience with mlops tools through Project based learning. The course is organized into six milestones focusing on different aspects of mlops, such as reproducibility, quality assurance, API development, containerization, and monitoring. The pilot run of the course received positive feedback, with all students successfully deploying their machine learning components to the cloud. The software Resurrection exercise helps learners understand software maintenance by building and testing historical software releases, fixing issues, and reflecting on the experience. The exercise allows learners to experience the challenges and joys of software maintenance, contributing to readability, intelligibility, and maintainability of program code and documentation. Teaching programming patterns can be done at the later part of an undergraduate course Students should have a background in compiling software and programming in a language like C Learning to write programs that are easy to understand and maintain is valuable for both students and professionals Made with HARPA AI

I want a picture with bob! and an Ether!

대단하신분🎉

I have a question would it be possible to have more than just User and Kernel space?

Look at my man yann. so inspirational...

Great work, wish the talk were longer

interesting.

Thank you for sharing. I can not wait to check this paper.