The end of the “Age of Data”? Enter the age of superhuman data and AI
Everything ends, many things start again
2024
The Informed Intuition Model: Blending Data with Gut Feel for Smarter Decisions
If you have been following my journey for a while, you’re probably aware of my fascination with decision-making frameworks. The line between success and failure, whether in businesses or individual...
Measuring and Mitigating Hallucinations in Large Language Models: A Multifaceted Approach
(Interestingly, this is the first time arXiV has declined a submission from me. I would give the editors kudos for finally taking their role seriously if it wasn’t because I suspect this is simply ...
Beyond Token Prediction: the post-Pretraining journey of modern LLMs
(This blog post, as most of my recent ones, is written with GPT-4 assistance and augmentation)
From Taskmasters to Trailblazers: The Challenge-Inspire Model of Leadership
The breakneck speed of innovation in the artificial intelligence sector has naturally steered my writing towards AI-centric themes on this blog. Regular readers, however, will recall that my passi...
2023
AI as the New Member of the Engineering Team: Crafting an End-to-End AI Application with AI
Exploring the AI-Driven Future of Software Development
Beyond Singular Intelligence: Exploring Multi-Agent Systems and Multi-LoRA in the Quest for AGI
In my view, the concept of Artificial General Intelligence (AGI) as it’s commonly understood might be a misnomer. Human intelligence itself is not ‘general’; it is inherently constrained by our sen...
My most epic runs
The Big Sur Marathon is considered the most beautiful marathon in the US and top ten in the world. However, it did not make it into my top 10 most epic runs
The Multimodal Generative AI Revolution is Here!
In the landscape of Generative AI (GenAI), we often find ourselves amazed at the rapidity and scale of advancements. GPT-4 stands as a shining example, pushing the boundaries of linguistic understa...
Beyond Prompt Engineering: The Multi-Layered Cake of GenAI Development
The layers of GenAI development
Mitigating LLM Hallucinations: a multifaceted approach
(I recently turned this guide into a paper. You can find it here)
On open source LLMs, take two: The three laws of GenAI model evolution
In this post I summarize the main advances in the area of LLM models, and particularly open source LLMs (including Falcon, LlaMa2, and Free Willy). I describe different leaderboards and what their ...
Prompt Engineering 201: Advanced methods and toolkits
DALL-E 2: An old professor with a notebook in his hand talking to a futuristic looking robot. 4k. Professional photo. Photorealistic
Some of my favorite Recsys resources and links
Introduction to recommender systems: Basics and classic techniques Beyond the basics Ranking Factorization machines Explore/exploit Full page optimization Co...
Towards Reinforcement Learning with AI Feedback (RLAIF). What open-sourced foundation models, instruction tuning, and other recent events mean for the future of AI
A few weeks back I shared my thoughts on how things were going to evolve in the race to build better/larger/smarter generative AI models, and particularly LLMs. Here is what I had to say:
Blueprints for recommender system architectures: 10th anniversary edition
Ten years ago, we published a post in the Netflix tech blog explaining our three-tier architectural approach to building recommender systems (see below). A lot has happened in the last 10 years in ...
2022: A Year in Books
I have always loved reading. For the past 3 years I have also gotten into audiobooks, which has increased my book consumption dramatically. I listen to many audiobooks, and then buy in paper those...
2022
ChatGPT: hype or revolution
“We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.” - Amara’s law
A new chapter
(Read my cofounder Neal Khosla’s version in Curai’s blog )
An unnecessarily long and personal recount of a 70.3 Ironman in Andorra
I have not written about my running/sports feats in a while. On the one hand I don’t think I have anything more interesting to say than what I said in my 10 Lessons I learned from running. On the o...
Migrating Medium and Blogger blogs to self-maintained Github Pages
In case you are reading this post in my now Medium mirror, you can access the result of this migration here.
The death of the stars: A brief primer on online user ratings
A few weeks back, Netflix announced that they were adding a two thumbs up option to give feedback on their content. At the same time I was being asked on Twitter why we had given up on 5 star ratin...
2021
On the “Usefulness” of the Netflix Prize
It has been over 10 years since the Netflix Prize finished, and I was not expecting to write a blog post about it at this point. However, just in the past couple of weeks I have found myself talkin...
The AI/ML FDA Plan
In case you are not familiar with it, the Food and Drugs Administration (FDA) is a federal agency of the Department of Health and Human Services. that is responsible for the control and supervision...
On Decoupling
You have probably heard of the butterfly effect: a distant butterfly flapping its wings can have a hard-to-predict effect on a future event. Everything seems to be connected in the world, and tech ...
2020
Scaling healthcare through AI
Curai’s mission is to “provide the world’s best healthcare to everyone”. In order to do so, we strive to scale the reach of an individual physician by creating tools that automate large portions of...
The 7+1 Habits of Highly Successful Leaders
I recently had a candidate I was interviewing to ask about my leadership style. I gave a short answer, but I was left thinking that I would have liked to continue talking about this for a much long...
2019
The year in AI: 2019 ML/AI advances recap
It has become somewhat of a tradition for me to do an end-of-year retrospective of advances in AI/ML (see last year’s round up for example), so here we go again! This year started with a big recogn...
Data as Prior/Innate knowledge for Deep Learning models
There is a long history of debate about how much of human knowledge is innate and how much is learned from experience/data. This is also known as the nature vs. nurture debate (See for example some...
2018
The year in AI/ML advances: 2018 roundup
It has become a sort of tradition for me to try to summarize ML advances at this time of the year (see here for my Quora answer last year, for example). As always, this summary will necessarily be ...
The Memo Culture
Some days back I was having a conversation about the importance of memos and documents in an agile/fast-moving startup company. The person I was talking with was surprised with my thoughts and aske...
Curai
I haven’t updated this channel in a while. I hope that by now you have already found out elsewhere that about a year ago I co-founded a company in the AI/Healthcare space: Curai. A few weeks back, ...
2016
Football or Futbol? Or why Deep Learning will not make other Machine Learning approaches obsolete
You have probably heard a lot about Deep Learning and how it is taking over the world in general, and the area of Machine Learning in particular. But, does that mean that Deep Learning will be the ...
Evil ads, ad blockers, and the end of Web diversity
Ad blockers might be good for the individual consumers that are using it, short term. However, longer term, if successful, ad blockers will shape the Internet in a way that will end up hurting cons...
How do you deal with strong difference of opinion among expert engineers in the team?
This is a great question. As a TLDR; the way to deal with strong differences between engineers is not necessarily to build consensus but to have a clear owner in charge of making the final decision.
Where am I?
There have recently been some articles (e.g. This list of influencers) that have pointed to this blog and lamented that I don’t update it regularly anymore. It is true. I now realize I should have ...
What’s the relationship between machine learning and data mining?
This is not an easy question because there is no common agreement on what “Data Mining” means. But, I am going to say that I disagree with the answer from Wikipedia that Yuvraj Singla points to. I ...
10 More lessons learned from building real-life Machine Learning systems — Part II
In my previous post, I introduced the ten new lessons and described the first five. Let’s directly dive into the final 5.
10 More lessons learned from building real-life Machine Learning systems — Part I
Over a year ago, following an original presentation at MLConf, I wrote a blog post entitled “10 Lessons Learned from building ML systems”. At that point, I was leading the Algorithms Engineering te...
2015
2015: My Year in Running
2015 has been a great year for me in many ways, but running has been particularly good. This year I ran 2000 miles (3200 km), compared to the 1850 last year. This is even more remarkable if I think...
2014
Ten Lessons Learned from Building (real-life impactful) Machine Learning Systems
(This is a blogpost version of a talk I gave at MLConf SF 11/14/2014. See below for original video and slides)
Introduction to Recommender Systems: A 4-hour lecture
A couple of weeks ago, I gave a 4 hour lecture on Recommender Systems at the 2014 Machine Learning Summer School at CMU. The school was organized by Alex Smola and Zico Kolter and, judging by the a...
Blog posts and Summer gigs
I have recently heard complaints that this blog is rather quiet lately. I agree. I have definitely been focused on publishing through other sources and have found little time to write interesting t...
2013
Recommendations as Personalized Learning to Rank
As I have explained in other publications such as the Netflix Techblog, ranking is a very important part of a Recommender System. Although the Netflix Prize focused on rating prediction, ranking is...
Tools for Data Processing @ Webscale
A couple of days ago, I attended the Analytics @Webscale workshop at Facebook. I found this workshop to be very interesting from a technical perspective. This conference was mostly organized by Fac...
10 Little lessons for life that I learned from running
(Sorry for allowing myself to depart from the usual geeky computer science algorithmic talk in this blog. I owed it to myself and my biggest hobby to write a post like this. I hope you bear with me...
2012
Recsys 2012: A long (and likely biased) summary
After a great week in beautiful and sunny Dublin (yes, sunny), it is time to look back and recap on the most interesting things that happened in the 2012 Recsys Conference. I have been attending th...
Netflix @ Recsys 2012
We are just a few days away from the 2012 ACM Recommender Systems Conference (#Recsys2012), that this year will take place in Dublin, Ireland. Over the years, Recsys has become my favorite conferen...
More data or better models?
The discussion of whether it is better to focus on building better algorithms or getting more data is by no means new. But, it is really catching on lately. This was one of the preferred discussion...
Beyond the 5 Stars Round up
Last week, I published a post on the Netflix tech blog. The post, entitled “Netflix Recommendations: Beyond the 5 stars” describes how recommendations have evolved at Netflix since the Netflix Priz...
2011
Recsys 2011 – Notes and Pointers
I found Recsys this year of very high quality in general. There were many good papers and presentations. The Industry track was also very high-quality, with very interesting talks from companies s...
The Recommender Problem & the Presentation Context
In the traditional formulation of the “Recommender Problem”, we have pairs of items and users and user feedback values for very few of those dyads. The problem is formulated as the finding of a uti...
Joining Netflix
Three weeks ago, I started to work for Netflix. Everything has moved so fast with so many things to do and learn that it seems like I have already been here for a much longer time!
Walk the Talk: On the Combination of Implicit and Explicit Feedback
Last week, Denis Parra presented our paper entitled “Walk the Talk: Analyzing the Relation between Implicit and Explicit Feedback for Preference Elicitation” at the UMAP conference. The paper won D...
Recommender Systems: We’re doing it (all) wrong
A few days back, there was an interesting post by Judy Robertson in the Communications of the ACM blog. The post, entitled “Stats: We’re doing it wrong“, builds upon a paper from last year’s CHI co...
The Science and the Magic of User Feedback
That was the main title of a series of talks I gave in different labs and companies during my recent California tour. In this presentation, I talked about many of our recent projects related to how...
Managing Research the Agile Way
I have discussed previously on this blog about how well the Scientific Method adapts to Agile approaches. These ideas also took me to an unfinished effort to draft an Agile Research Manifesto. Howe...
On Trust Networks and Gamification. Or How Quora can overcome its Hype and embrace long-term Success
If you are reading this blog I am pretty sure that you know quite a lot about Quora by now. If not, you should sign on and try it a bit before you continue reading the post.
2010
Did you prepare your talk?
I don’t consider myself to be a great presenter. As a matter of fact, every time I finish a presentation, I find myself thinking about how many things I screwed up and could have done much better. ...