วีดีโอ

Is there a way to find out what is the paper she chose to present for this program? Trying to get an idea how recent the publication paper should be :)

This guy is amazing and he knows how the catch the viewers.. Respect to Andrew..

@Edward Robinson First of all, thank you for sharing such great knowledge. Second, I have a question. I have Pakistan's laws dataset, and thousands of laws are in different categories. How would I build RAG architecture to provide the best answer to user queries relevant to those laws?

Thank you to the entire team for this insightful presentation on the Multi-Agent Aya for Workforce project. The innovative approach to addressing communication barriers in distributed enterprises through a multilingual, context-aware chat application is truly commendable. The integration of AI to enhance efficiency and streamline workplace communication is a significant step forward. We look forward to seeing the future developments and integrations you have planned. Excellent work!

How can I join these live in future ?

@ Stephen Colbert ❤

this is going viral

💯🚴🏃🏃🤸🙏

39:50

k-means is poor man's analysis. It has little to no statistical reasoning for clustering. Works off heuristics 😓

Great video. Very insightful. Appreciate you guys sharing these insights on a solution that's in in production.

He doesn't look like a Salamanca. Lalo maybe?

Ahmad ❤ So proud of you ❤

Thanks for this Cohere. Looking forward

who can apply for this programme? can newbies in ML like me are eligible? I am a undergrad student pursuing btech cse and having great interest in AI/ML

Very well put discussion and very on-point. I'd like to recommend a follow up: deployment strategy from POV of enterprises and foundation model developers. Thanks.

Inspiring Journey. I'm going to apply for this year. I hope it works out :)

instead of holding everyone's hand, why not simply teach people to think critically. do you trust everyone you meet? then why should it be any different with LLM or any new shiny technology? by imposing constraints, you create cost on everyone, like California's new legislation on AI, that even Pelosi opposes. wouldn't it be better to teach people to think than not think?

Congratulations on the achievement sir Ahmad!

That was great.

I missed the info session on 16th Aug. Could you share the live session with us? It would be helpful to share it only for a limited time.

amazing!

Бро настолько попали в Евразию, что молились всем Скандинавским богам, лишь бы не ткнуть в Россию я ущемлён и требую личных извинений

Бро настолько попали в Евразию, что молились всем Скандинавским богам, лишь бы не ткнуть в Россию я ущемлён и требую личных извинений

Such an insightful conversation

Nice any new model?

Vijay, I wonder the amount of AI domain knowledge you are holding to give such a wonderful talk about AI and its integration to Workflow/Service Automation. Really appreciate.

Quite clear!

Absolutely fantastic explanation ! Thank you very much for this video ! Particularly neat the explanation for sampling as a 2 step process : Going to predicted x0 and noising the prediction with the forward diffusion process one step closer to the original x0. All training objectives and sampling steps made total sense and felt the natural thing to do.

Nice.

Great talk! Thank you soooo much

Ahmet hocam başarılar, gercekten cok guzel bir proje icerisindesiniz !!!

I was going to say, stop with the AI, and employ real people instead. Ground breaking idea, I know! Unless you want a world where people don't feel motivated to learn to do anything, after all an AI could do in one minute what you could do in weeks, with years of experience behind you, yeah, why even bother? But the recording of sound is so bad, that maybe only AI could salvage it. If this was intentional, then well done.

The sound is terrible!

Command R+GraphRAG finetune coming soon

8:31 I think there is an abstract layer people are applying reasoning to and taking for grandeur. Even though we think only moving one spot away per move in the direction of weigh point is common sense and reasoning is the opposite of the reason why the model are able to excel at higher dimensional vector search and semantic meaning. The system never learns that the correct answer is X it only understand the relation between positive affirmation and the incorrect answers. Not am I saying the intellect score is a positive attitude but as for the the end-use right answer - might be wrong. General AI is Narrow AIs. 30:11 See I don't think they have noticed non determinism but just found a glitch in the reward mechanism through General Habit by exploiting human characteristics that the error margin is lost between all noise. It probably still needs A* data sets but with deterministic entropy in the weights is what I am saying;. Just because the models excel at whatever benchmarks doesn't demonstrate actual comprehension is the point.

perfect!

Las explicaciones de Luis siempre son Geniales. Gracias!

It is a great session. Do you mind sharing the slide?

Does this mean, we can start the process in say, Swahili and then convert to English, so as not to lose the particular characteristics of Swahili which happens when one starts with English?

1st

I was considering training a model on deliberate problem-solving with multiple Fermi problems, sadly there I could not find good data

good

This was a great session. Thanks.

15:00

No Fear.

Would be awesome if you could link papers!

Consider me convinced

awesome

thank u so much sir, greetings from turkiye