Amol Borkar

A study in perspectives on AV safety: a Waymo robotaxi approaches an intersection in San Francisco. Some other driver is doing "really fast donuts" in the intersection, and has been doing so long enough that the robotaxi rider looks up because of the smell of burning rubber entering the cabin. Robotaxi enters the intersection, then stops. The other driver halts and narrowly misses hitting the robotaxi (by perhaps a foot). The rider: "I can trust the Waymo, but I can't trust this driver". States that the other driver was upset but the robotaxi rider "shrugged, and pointed to the invisible driver." What ya' gonna do? It's a Waymo. "Still going to use Waymo." Presumed Waymo response to regulators and media: The human driver doing donuts was irresponsible. If there had been a crash it would have been the other driver's fault. (If I got this wrong would be happy to be corrected by Waymo. They have not commented as far as I know.) A broader safety engineering response: This is not the first time a robotaxi has encountered another car doing donuts in an intersection. In fact, a Cruise robotaxi was hit in a similar situation in March 2023 (SGO 30412-4960). From the description it seems there was plenty of warning this was happening. Why did the robotaxi not stop well clear of the intersection as a prudent human driver would have? Is Waymo monitoring SGO data and learning from what has been reported by other companies? Has Waymo included "other car doing donuts" in its hazard log? Did Waymo decide to ignore mitigation of this hazard for some reason? Will that change now that this has happened to Waymo instead of just Cruise? I'm glad in this case there was no injury, but there certainly could have been but for chance from a rare circumstance that is already known to have caused a crash. Being saved by the “grace of the Gods” is not a safety argument. I hope these questions are being asked and resolved by Waymo. But we won't know until the next donut incident. Primary Source: https://lnkd.in/eFHEMVFX Article: https://lnkd.in/eFVpfjec

65

14 Comments

Tonight, I stepped into the future 🚗 24 min. fully autonomous Waymo in SF. From complex turns to city traffic, it handled every move flawlessly. Safety. Affordability. Innovation. 🤯 📍Here’s what caught my attention: 1️⃣ Safety First – I felt secure throughout; the technology was tuned to anticipate every turn and stop. 2️⃣ Affordable Access – Costing less than my other rideshare options, it hinted at a future where AVs make mobility more accessible. 3️⃣ Invested Future – Alphabet’s $5.6 billion investment isn’t just about tech. It’s about making transportation work for people. 🌊 The potential is here and now. Autonomous vehicles are set to redefine what it means to travel: safely, affordably, and sustainably. Imagine a world where transportation is truly built for people. ❓Would you ride in a fully autonomous vehicle? Follow Elise Victor PhD

5

4 Comments

Tesla is one tech topic that has unconditional support. No matter how hard you point out some defects in it, people just love red cars. When I first began sharing my Tesla observations, people have been telling me that the info must be coming from Elon Musk haters. They might be right, as Tesla has been experiencing resentment after the hostile and rough takeover of Twitter. But, there are some simple truths that are hard to miss. Like pointing out the fact that Tesla experiences routine problems that have been solved by carmakers for decades now. Like the front wheel falling off, this and that common part not working as intended, etc. A dear friend shared with me that the Tesla he drove had water leaking from the roof when driving in the rain which left his documents wet. I mean, this is a solved problem. I am not encouraged to buy the bleeding edge of cars for my patience to be massacred by routine, casual, ordinary, dumb issues. Even people who encountered some Tesla issues are quick to defend Musk's cow. Maybe this issue that one other person reported was not exactly like that. Maybe this did not happen or it happened this other way. How did we get to this point? Maybe because they popularised novel features the right way? Maybe a lot of engineering went into it? Maybe the feeling is great when driving? Maybe. There are lots of reasons to love Tesla, but, when it's flawed, it's flawed. Love is blind. Red love blinds.

1

Checkout our latest report on how just eight SN40L chips run 1 Trillion parameter model at blazing fast speed https://lnkd.in/grwdFVAi. It was great to collaborate with some of the most talented folks in this area. Here is the link to the demo operating on similar principles https://fast.snova.ai/

28

Reflections on my first-ever self-driving Uber (Waymo) experience... When I told my family and friends about my first Waymo, autonomous vehicle experience, I was surprised by the reactions. "You are so brave." "Weren't you scared?" "I have driven in SF, that is insane." "I never even knew that was a thing." "That sounds so risky." "Only in California." I reacted by saying that these cars were all over the city, and they seemed very safe, reliable, and even cool. Most importantly, they really seemed to work. In fact, I recall watching people exit the vehicles, and they all looked happy, excited, and really glad they took the risk. I began to reflect on how this experience equates to pharmaceutical companies first choosing to use Mobile Locker over the safer, more comfortable, and more visible choice(s). After all, the small, lesser-known experience is a bit riskier, right? But in the end, it turns out we, too, are "all over the city," with nearly 60 satisfied pharma/med device clients. And yes, I think most would say we are safe, reliable, and...(well, okay, maybe not necessarily cool.) Finally, most every one of our clients is happy, excited, and really glad that they took the risk. Why...because Mobile Locker really works. So if you are seeking to provide more impact and value to your field and to the HCPs you are targeting with your digital messaging... Consider riding with us here at Mobile Locker. #pharmaindustry #meddevices #hcpengagement

18

2 Comments

In a recent interview, NVIDIA's CEO, Jensen Huang highlighted Tesla's leading position in self-driving technology. I have been using Tesla's #ModelY in Full Self-Driving mode for over a month now, and I couldn't agree more with the CEO's assessment. The experience truly feels like something out of a sci-fi movie. On highways, #FSD excels at lane changes, safe distance maintenance, and merges. Its navigation through complex interchanges and adaptability to varying traffic conditions impressed me, making long drives more relaxing and enjoyable. In neighborhood settings, #FSD's attention to detail in recognizing stop signs, traffic lights, pedestrians, and cyclists stood out. It navigated turns and intersections smoothly, ensuring a safe driving experience even on busy streets. The technology's advanced capabilities in navigation, lane changes, and automated parking have significantly enhanced my driving experience. The seamless integration with real-time data and intuitive interface instilled a sense of confidence and safety. Beyond the autonomy, driving a Tesla is exhilarating in itself. The instant torque and rapid acceleration deliver a thrilling experience, with a pickup comparable to racing cars. This powerful performance, combined with cutting-edge technology, makes every drive an adventure While there is room for improvement, Tesla's #FSD undeniably represents a substantial advancement in automotive technology. Highly recommend exploring this innovative and transformative technology! #Tesla #FSD #Innovation #AutonomousDriving, #ElonMusk

13

1 Comment

Check out our new paper on "Joint Semantic Knowledge Distillation and Masked Acoustic Modeling for Full-band Speech Restoration with Improved Intelligibility"! We introduce a generative model addressing denoising, dereverberation, bandwidth extension, and declipping, producing high-quality full-band (44.1 kHz) speech. Our approach involves masked audio token modeling with a language model. Notably, we propose self-supervised semantic knowledge distillation, which significantly enhances the speech intelligibility without the need for transcribed data nor increasing model size and inference time. Find the paper and samples at the provided link!

41

"Analog design remains stubbornly specialized" From a simulation perspective it isn't specialized, but the analog guys seem to like buying specialized SPICE simulators over using Verilog-AMS simulators and just specializing the models. If you want your analog engineers to be productive, don't let them use archaic commercial SPICE simulators. Better still, hire me and I'll build you an AI to do the work ;-) https://lnkd.in/gkPzn4xE

8

Faculty friends: We've recently opened the NVIDIA Academic Grant applications. The "Graphics & Vision" track covers lots of XR and Spatial computing topics: == Augmented Reality and Virtual Reality == > Digital twins > Human-computer interaction for immersive visualization > Immersive analytics > AI for immersive AR / VR > High-end visualization systems (multi-node, large scale visualization research systems) > 3D scene composition and content capture

56

1 Comment

Just wrapped up the latest edition of The Compiler, and it's brimming with tech insights that could shape your engineering strategy 💡 Here's a snapshot: • OpenAI's hardware ambitions: Potential Cerebras acquisition signals the rise of specialized AI chips • Record-breaking $35.5B invested in AI startups in H1 2024 alone 📈 • Polars vs. Pandas: Performance showdown for data engineers • Google's innovative C++ memory safety retrofitting • Advanced metaprogramming: C++ templates vs. Lisp macros • Scaling Django for high-load scenarios • Gliimly: New language for efficient web services • Pex: Simplifying Python packaging and deployment • Go-taskflow: Enhancing concurrency in Go • Google Docs now with AI-powered image generation 🖼️ Curious to dive deeper? Check out the full newsletter here: https://lnkd.in/gVD_SQHh What caught your eye? Let's discuss in the comments! 🤔

3

It is good to see vendors racing to introduce the next-gen HBM memories at a 2-2.5-year cadence. I hear rumors about production-ready HBM4 parts in 2026… With the growing demand for memory bandwidth from GPUs and AI accelerators, these announcements are encouraging for the AI community. However, the downside of this increased demand for the highest bandwidth HBMs is that, with limited production capacity, the life cycles of previous-generation parts are shrinking. The HBMs typically have a life cycle of 4-5 years. If customers need older-generation parts for longer, they must purchase them in bulk (last-time buys) and stock them up. But the shelf life of standalone HBM parts is only around 2 years! After that, they would have to be assembled in the package with the core dies to extend the shelf life by another few years. These shorter life cycles are hurting other non-AI customers, such as networking system vendors. Many switching silicon built with HBM2/2E parts could be used for products spanning more than 10+ years, as not every customer needs the highest bandwidth switches. Since each generation of HBM is not pin-comparable with previous ones, at EOL, they are either forced to re-spin their core dies to interface with next-gen HBM parts or shorten the life cycles for the systems they release. I hope the JEDEC community considers interoperability between different generations of HBM. This would help all applications using HBM. I agree that there are challenges; it is hard to double the bandwidth with the same set of interfaces and pinouts. Should they consider moving to other standard die-to-die interfaces like UCIE that guarantee one generation of interoperability? Any thoughts? #sharadayeluriblogs #routers #ethernetswitch #junipernetworks #highperformancecomputing #gpus

109

11 Comments

Great webinar from my colleagues Tianyi Zhu and Maggie Oltarzewski on #simulink modeling, simulation, and testing! Great intro for those new to Simulink, but also a good refresher for those who need it: https://lnkd.in/eg8qgVec #matlab

11

Its sometimes hard to believe that we live in an era where driverless vehicles aren't just a vision anymore but a reality. Waymo driverless cabs for example, are transforming the San Francisco streets forever! It was a wonderful experience working on an application of autonomous vehicles as part of my final project in the course CS267: Applications of Parallel Computing at the University of California, Berkeley. Our team, comprising Terence N., Zeid Solh, and myself, focused on parallelizing stereo vision for driverless applications. Stereo vision is crucial for self-driving cars as it builds a 3D understanding of the environment by using two cameras. However, real-time processing of these images is computationally expensive, posing a significant challenge for autonomous vehicles. We tackled this by exploring various parallel computing techniques to improve the speed of calculating the Disparity Map algorithm. We used popular techniques such as #SIMD (Single Instruction, Multiple Data), #OpenMP (multi-threading), and CUDA (NVIDIA GPU acceleration), and benchmarked the performance. By distributing the computational load across multiple cores and processors, we achieved substantial performance improvements, enabling real-time processing of high-resolution images and videos. The results were promising! Utilizing CUDA, we saw significant speedups, demonstrating the potential for efficient, real-time depth perception in autonomous systems. This project signified the importance of parallel computing in advancing autonomous driving technologies. I'm incredibly proud of our team's hard work and the insights we gained through this project. Check out our GitHub repo: https://lnkd.in/gYHrEU9z for more details on our implementation! #UCberkeley #CS267 #ParallelComputing #StereoVision #AutonomousDriving #DeepLearning #ComputerVision

49

3 Comments

I've mentioned this to #investors outside of sports and entertainment. The adoption curve for technology in sports is a hockey stick. Years of pilots in fast paced, high stress environments, but if the application is scoped out properly with the right forces behind it can lead to BIG wins in terms of adoption. I don't know much about AiFi Inc. but this looks to be an awesome example on the adoption for checkout free applications. Well done. For #roboticsINsports check out Hop Robotics

18

1 Comment

I'm advocating for using markdown as a simple file format to recording and replaying LLM sessions, specifically using bird tracks before prompts. There are lots of ways to render markdown. HaverScript uses this format already for the echo mode. Thoughts? https://lnkd.in/gbfN7mBf

8

1 Comment