Tom Rochette

I just released the April edition of my AI & the Web newsletter. Nathaniel Simard, founder & CEO of Tracel Technologies and inventor of the burn ML framework, spoke with me about how he got into ML, burn, Tracel and the future of AI. WebGPU, the successor of WebGL, will soon also be supported in Firefox and Safari. Small, powerful LLMs running right in the browser is not a distant future, but are the next big thing in cost-effective and privacy-friendly AI. Google is working hard on better WASM support in the browser that makes ML models running fast even on low-end devices without GPU. With Llama 3 by AI at Meta and Phi 3 by Microsoft AI we got some exciting new, small modells. Llama 3 was already ported to run in the browser and we'll see Phi 3 soon as well. MusicGen has been ported to the browser in transformers.js by Hugging Face engineers and developed an online Jukebox running right in the browser. #webgpu #webml #ai #ml #llama3 #wasm https://lnkd.in/drJCY7Ew

19

2 Comments

Well NVIDIA is now worth more than the GDP of Canada and in fact, more than 14 of the G20. So what is behind this insane growth and valuation? 1. Exponential requirements for growth. If we look at "GPT5" the needs will be far more than GPT4. From GPT2 to GPT4 OpenAI progressively added orders of magnitude more computing power to calculate the models. Rough estimates are GPT4 used 3,000 to 10,000x the compute power compared to GPT2 2. Data centers aren't nearly big enough...by a lot. GPT4 benefitted from having a lot of excess compute in existing data centers. But the next models will need compute thousands of times bigger. Think $100b or $1T data centers. 3. Google, Microsoft, Amazon are investing billions (>10b each) to grow the data centers to meet the demands. And now I'll let you guess who is supplying those chipsets.... If you are looking for some details on this, Leopold Aschenbrenner of situational-awarness.ai wrote an excellent article on calculating computational requirements. But hey, keep your head up Canada, at least we've got a team in the NHL playoffs. https://lnkd.in/gadhkfxf

3

5 Comments

👀 If you look closely, you’ll notice a significant shift in Vancouver’s software-based startup scene. Many developer-centric startups are being built right here in Vancouver. Defang Labs is one of them. At Defang Labs, we’re building something developers love. Imagine deploying apps in the cloud in minutes, not hours. This is happening right here in Vancouver and I’m excited to be part of it. #VancouverTech #StartupScene #DevFirst #devtools #devex #developerexperience #defang

85

6 Comments

Just another day, where the fantastic team at ICP HUB Canada & US and Quantum Leap Labs is educating and empowering the world, one round table at a time😋 Truly, it was a delight curating this epic roundtable on the Topic - Building Projects from Zero to DAO on the ICP Blockchain at the Elevate Festival alongside one of Quantum Leap Labs' top founders, Sam Drissi, CEO at bIPQuantum, a revolutionary blockchain based solution changing the way we deal with Intellectual Property. (PS: You can clearly see his dedication in creating IPs out of random moments through his passion for photography. Just kidding👻) The intention behind doing the roundtable was to tap into startups that are building sustainable solutions, and to see if ICP as a technology could be used to optimise these solutions. And we did! We were fortunate to meet with great and innovative minds that wanted to utilise technology to its fullest. Huge shoutout to Ewelina E. Pankowska, Lhallia Bouima, Kavyne Bent, Devender (Dev) Butani and ofcourse Sam Drissi for turning the table into a brainstorm fiesta! But it wasn't exactly like that when we started it. Being in the industry for quiet sometime, working first-hand with web2 products and understanding the technological advancements of the web3 world, honestly gave me an unfair advantage. I was dealing with CEOs here, and using fancy web3 jargons was blocking the true purpose and power of blockchain. Which is why, I gave one look to Sam, and we hit the brakes on our conversation, going back to the inception and breaking things down to bits. We weren't afraid of talking about other blockchains and the amazing work they have done so far, because, at ICP — We believe in true "interoperability" which can only be achieved through collaboration. And so we did talk about how it all started, turning complexity in to clarity which lit the room up with "Aha" moments. My favorite part about such audience is the "Wait, you can actually do that?" moments as you dive deeper into the conversation. And boy, we had plenty! - when we explored ICP's scalability, robustness and the power of on-chain governance to make truly decentralized decisions, digging into how these can transform the needle of technological products. And guess what the outcome was? I'm lining up calls with their tech teams this week to convert these "deep-dive" conversations into real collaborations! A big shoutout to Lauren Crerar and team Elevate for their wonderful support, ever since the first call! Thank you for accommodating us in such short time and for bringing such quality crowd under the same roof. More power to y'all!🚀 And to Javier Arroyo Ferrer, thank you for believing in me, for giving me to the opportunity to shine each time and for being a great teacher! Forever grateful and indebted :) Here's a little sneak peak from the day, cheers!✨

112

6 Comments

Recently I interviewed Nathaniel Simard, founder & CEO of Tracel Technologies and inventor of the burn ML framework: Jan: Nathaniel, we're excited to have you as our guest today! To kick off our conversation, could you please introduce yourself and share a bit about your background and what got you into ML? Nathaniel: Sure, I'm the creator of Burn, a deep learning framework written in Rust, and the founder of Tracel AI. I started coding in the first year of university, where I was studying mechanical engineering, but I quickly switched to software engineering after, since I instantly fell in love with programming. Then I explored different facets of the field from backend to frontend development, and I decided to start my career as a consultant focused on software quality. After some time, I wanted to go deeper into AI, since I was always interested in the process of learning, so I enrolled for a master's degree at MILA. J: According to Github you started to work on burn in summer 2022. Since then has already earned over 7000 stars on GH. What was your initial motivation to develop a new ML framework? N: I always had a side project going on, for fun mostly and to learn new things. I wanted to explore asynchronous and sparse neural network architectures, where each sub-network can learn and interact with other sub-networks asynchronously and independently. I wasn't able to actually create something useful because I needed fine control over the gradients and the concurrency primitives, which is not easily done with Python and PyTorch. At the same time, I was working on machine translation models at my current job, and it was quite painful to put models into production. I decided to switch my side project to a new deep learning framework, with more flexibility regarding gradients and concurrency primitives as well as being more reliable and easier to deploy on any system. J: Amazing to see that this was born out of a side project! I feel the struggle with concurrency in Python and this is something where Rust really shines. What are some of the other key features that makes burn special? N: I think there are two things that really set Burn apart. First, almost all neural network structures are generic over the backend. The goal is that you can ship your model with almost no dependency, and anybody can run it on their hardware with the most appropriate backend, even embedded devices without an operating system. Second, Burn really tries to push the boundaries of what is possible in terms of performance and flexibility. It offers a fully eager API, but also operation fusion and other optimizations that are normally only found in static graph frameworks. The objective is that you don't have to choose between portability, flexibility, and performance; you can have it all! To read the rest of the interview with Nathaniel's plans for burn & trends in AI here: https://lnkd.in/dKv9ztMm #ml #ai

3

🚀 Excited to Share: C-Law AI is Almost Ready for Beta Testing! ⚖️ I’m thrilled to give you a sneak peek of C-Law AI, an all-in-one platform designed to simplify legal processes and make them more accessible for all Canadians. ✨ Here’s a Sneak Peek of Our Features: ✅ Legal Chat: Instant answers to your legal questions powered by AI. ✅ Document Check: Upload and verify your legal documents with advanced AI-driven analysis. ✅ Document Builder: Effortlessly create professional legal documents with our intelligent builder. While C-Law AI isn’t live yet, we’re gearing up for beta testing, and I couldn’t be more excited to share it with you all soon! 💡 Want to be part of the journey? Drop me a message if you’d like early access or want to help shape the platform during beta testing. Stay tuned for more updates! #LegalTech #AI #Innovation #BetaTesting #CLawAI #Canada #CanadianLaw

6

Checkout this demo by my co-founder Dov Amihod showing how he connected all his tools from Notion, Slack, Github, Calendar and more to AI agents in Datagrid AI to get stuff done!

20

Woo! Anthropic has finally made Claude/Claude Pro available to Canadians. Very happy to have direct access to it for testing. I also like the fact that they explicitly say they don't train on the data from interactions. For reference; "By default (and just like our beta version of Claude.ai), we will not use your prompts and conversations from Claude Pro to train our models. There are two instances in which we may use your prompts and conversations to train our models: (1) if you give us explicit permission by submitting feedback through the thumbs up/down feature or by reaching out to us with a request, and (2) where your prompts and conversations are flagged for trust and safety review, we may use or analyze those conversations to improve our ability to detect and enforce Usage Policy violations, including to train Trust and Safety classifiers in order to make our services safer. Only a limited number of staff members have access to conversation data and they will only access this data for explicit business purposes." #anthropic #ai #claude #claudepro #Canada #access

2

“Who in the audience already thinks that AI is conscious?” I asked a Vancouver AI crowd last night. Check out the first 3 minutes of my presentation below. Full video on YouTube: https://lnkd.in/ge-dc3wa Thrilled to be able to have conversations about consciousness in AI. Thanks to Kris Krüg for the opportunity! #ConsciousAI #YouAreConscious #VancouverAI

101

4 Comments

One of the hard parts about being a data scientist is learning what "going to production" means. You will get questions during interviews like "what have you taken to production", "is that production level code"...wut? Going to production means many things to many people. Well here is a great bootcamp demo run by 🎧 Eric Riddoch whom I would trust to teach you his version (and a solid definition at that) about what it means to go to production using Python for ML type projects. 🎧 Eric Riddoch has some serious skills in this area and is a trusted colleague. Nope, I make no money posting this on his behalf, I just trust his skills enough to be willing to shout it out to anyone within earshot. Give it a look and if you feel its for you, give it a try. Eric can give you more details about his full bootcamp as well.

21

2 Comments

Last week, Steven Astorino challenged Tanmay Bakshi and I to demonstrate building and deploying an application from scratch at the IBM Canada Townhall. We made this possible by leveraging watsonx Code Assistant to accelerate our workflow, allowing us to focus on the inventive components of software development. Since the Grammy Awards are coming up, we decided to build a polling platform where the audience votes for nominees in each of the major categories so we could determine IBM Canada's predictions for the outcomes. watsonx Code Assistant helped us write boilerplate code, SQL queries, data structure transformation logic, and more! As a result, we had a fully deployed application in ~15 minutes and collected over 1400 votes from attendees. Turns out, IBMers share my opinion that Taylor Swift will win Album of the Year 😊

187

8 Comments

𝐌𝐢𝐜𝐫𝐨𝐬𝐨𝐟𝐭 𝐕𝐚𝐧𝐜𝐨𝐮𝐯𝐞𝐫 𝐉𝐨𝐛𝐬 𝐚𝐧𝐝 𝐂𝐚𝐫𝐞𝐞𝐫𝐬 - Latest roles October 5th We've got several new job openings across various teams in Microsoft Vancouver this week. Interested? Apply now using the links below or share this post with someone who might be! (Please note - I'm not the hiring manager for these toles) I get asked to look at resumes a lot, but a generic resume doesn’t help much, especially when applying to Microsoft. You want your resume to match the job you’re going for. Make sure it speaks to the role and highlights what’s most important. And to be honest, I’m not an engineer, so I might not understand all the technical details. You’ll get better feedback from someone who knows the job you’re after. 𝐒𝐨𝐟𝐭𝐰𝐚𝐫𝐞 𝐃𝐞𝐬𝐢𝐠𝐧 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫  https://lnkd.in/guwkZruD  https://lnkd.in/gaaJcYX4 𝐒𝐨𝐟𝐭𝐰𝐚𝐫𝐞 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫 https://lnkd.in/gtZwYFFh  https://lnkd.in/gqZgPtHd  https://lnkd.in/gZ8b-NzD  https://lnkd.in/gNb8ZT7v  https://lnkd.in/gvVkfsVe  https://lnkd.in/gG-iZ3WK  https://lnkd.in/gREpaSaT  https://lnkd.in/gTpmPVid 𝐒𝐞𝐜𝐮𝐫𝐢𝐭𝐲 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫 https://lnkd.in/g3ckHMzX 𝐏𝐫𝐨𝐝𝐮𝐜𝐭 𝐌𝐚𝐧𝐚𝐠𝐞𝐫 https://lnkd.in/gYp_sW-Q 𝐏𝐫𝐨𝐝𝐮𝐜𝐭 𝐌𝐚𝐧𝐚𝐠𝐞𝐫: 𝐈𝐧𝐭𝐞𝐫𝐧 𝐎𝐩𝐩𝐨𝐫𝐭𝐮𝐧𝐢𝐭𝐢𝐞𝐬 https://lnkd.in/gBx-QbPi 𝐀𝐧𝐢𝐦𝐚𝐭𝐢𝐨𝐧 𝐃𝐢𝐫𝐞𝐜𝐭𝐨𝐫 - 𝐆𝐞𝐚𝐫𝐬 𝐨𝐟 𝐖𝐚𝐫 - 𝐓𝐡𝐞 𝐂𝐨𝐚𝐥𝐢𝐭𝐢𝐨𝐧 https://lnkd.in/gFXHbK-s 𝐒𝐨𝐟𝐭𝐰𝐚𝐫𝐞 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫 - 𝐗𝐛𝐨𝐱 𝐆𝐫𝐚𝐩𝐡𝐢𝐜𝐬 https://lnkd.in/g_eKrGFX 𝐒𝐨𝐟𝐭𝐰𝐚𝐫𝐞 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫 - 𝐆𝐚𝐦𝐢𝐧𝐠 𝐀𝐈 https://lnkd.in/g2JfHFm9  https://lnkd.in/g9GrTeVq 𝐏𝐫𝐨𝐝𝐮𝐜𝐭𝐢𝐯𝐢𝐭𝐲 & 𝐀𝐈 𝐓𝐞𝐜𝐡𝐧𝐨𝐥𝐨𝐠𝐲 𝐒𝐩𝐞𝐜𝐢𝐚𝐥𝐢𝐬𝐭 https://lnkd.in/gsNeVM28 𝐂𝐥𝐨𝐮𝐝 𝐒𝐨𝐥𝐮𝐭𝐢𝐨𝐧 𝐀𝐫𝐜𝐡𝐢𝐭𝐞𝐜𝐭 https://lnkd.in/gnMF2sqn #Microsoft #MicrosoftVancouver #Vancouver #TechJobs #Hiring #SoftwareEngineering #JobSearch #CareerOpportunities #MSV

43

2 Comments

Thanks to #NeurIPS2024 taking place in #Vancouver this year, I’ve had the chance to attend incredible events and learn from some really smart people. Here’s a short clip from the Infer event earlier this week, where Shawn swyx W gave a fantastic overview of the current AI landscape. #AICommunity #TechEvents

36

We're thrilled to have Patrick Halina, Engineering Manager at Pinterest, join us at the 8th Annual Toronto Machine Learning Society (TMLS) Summit! Since the early days of the internet, website scraping has been a common technique for building datasets. However, the advent of Large Language Models (LLMs) has significantly changed it. Patrick's talk will delve into Pinterest's learnings from applying cutting-edge LLMs to webpage comprehension and information extraction. He'll share insights from processing over 1 billion web pages daily, including: ✅ An overview of the latest systems for webpage data extraction ✅ Pinterest's experience testing various approaches, from GPT and open-source LLMs to simpler models ✅ Research findings on their internal web extraction method See his full abstract here: https://lnkd.in/gDCHCUha #TMLS2024 #MachineLearning #LLMs #WebScraping #DataExtraction

20

1 Comment

🚀 Excited to be at ZK Hack Montréal! Over the next three days, I'll be diving deep into zero-knowledge proofs, collaborating with top developers, and exploring innovative uses of ZK technology. This event is not just about coding but also learning from the best, networking with fellow ZK enthusiasts, and pushing the boundaries of what's possible in blockchain and privacy-preserving technologies. 📸 Here's a snapshot from the event. If you're around, let's connect! #ZKHackMontreal #Blockchain #ZeroKnowledgeProofs #TechEvent #Networking

42

2 Comments

Good Canadian Kid Does and Sings AI Nick Frosst, co-founder of AI company Cohere, simultaneously leads a successful indie band called Good Kid. An unexpected blend of technology and music! Song recommendation: Listen to "Nomu" by Good Kid. The song includes interesting references to technological topics such as syntax, semantics, and lost language - concepts related to the world of AI. https://lnkd.in/dKBpKx5F This shows how the band members' technological background influences their musical creation. "The syntax skips, semantics slip And what was left a growing cleft of language lost, tokens unknown" To view the full lyrics: https://lnkd.in/d-78VdXr A moment of nostalgia: "How do you convert PDF to JPG?" by the (Israeli) band Umlala? What a classic! Today, we'd ask an LLM how to do that ;-) https://lnkd.in/dyMiq9cv What musical association comes to your mind?

1