Gen AI, Data Ecosystems, and Robots: from barcodes to apis

What do VCs think about it? and what's the impact of CrowdStrike on robotics?

CrowdStrike brings the “Blue Screen of Death” to the Las Vegas Sphere

Highlighting the vulnerabilities in all of our tech stacks, a CrowdStrike software update caused a global IT outage, leading to widespread disruptions across industries such as aviation, banking, and healthcare by causing Windows systems to crash with the "Blue Screen of Death."

No wait! Most robots use Linux.

Creating the Future of Robotics

Jeff Frick did an amazing job wrangling our panel at the Create The Future Summit. He covered a lot of ground in our very short 30 minutes. And we had robust discussion about quite a few points.

Robotics is the technology of the 21st century, that I believe is going to have the biggest lasting impact. Robots are physically in the world with us - Andra Keay

There’s a reason people are getting excited about humanoid robots. And that’s that they can go where people go. - Michael Harries

The rule in robotics, as well as any design, is form follows finance.  Ultimately the best robots are going to be the ones that are invisible. They do a function, integrated into the fabric of our daily lives. - Scott Summit

The Future of Robotics Fireside Chat - Create The Future Summit 2024, Michael Harries, Andra Keay, & Scott Summit,

Lessons for robotics from Walmart’s Rise to Dominance Through Their Data Ecosystem

An effective data ecosystem allows for seamless data flow, enabling organizations to derive meaningful insights and power advanced applications like generative AI.

Figure from HBR

Before the Gen AI revolution, we need the API revolution

To make sure all of the inventory data was received, Walmart developed its own satellite network in the mid-1980s, connecting all stores to its central computer system. This immediate sharing of sales data led to clearer insights about what was happening across the stores. What to stock, where to stock it, and when to stock it became simpler.

Walmart’s data flows and connections. Figure from EJLawless

The parallels with Walmart's data-driven transformation of retail are striking. Just as barcodes and satellite networks reshaped the competitive landscape of retail, the emerging Generative AI data ecosystem has the potential to redefine the nature of enterprise competition. The winners in this new era will be those who can effectively harness the power of AI to create more agile, responsive, and insight-driven organizations.

However, the true potential of Generative AI remains locked behind the challenges of building robust data ecosystems. The race is on to develop the APIs, processing tools, and storage solutions that will unleash the full capabilities of this technology. As these pieces fall into place, we can expect to see a rapid acceleration in the adoption and impact of AI across all facets of business.

Although this article focuses on GenAI, it relates equally well to the role of robots as data collection devices. Read more by clicking on the substack above by EJLawless

Very Useful Stuff

If you are a (VC, PE, Grant) funded startup looking to drive pipeline, conduct customer discovery, find of the voice of the customer, or a mixture of these - our team is just the right mixture of nimble and structured.

I’ve seen Fulcrum Sales and Marketing deliver a full pipeline of qualified customers to SBIR startups embarking on fundraising. And they do it affordably. They’re a startup secret weapon and superpower. :)

RoboCup 2024: All the Results

Humanoid RoboCup Results:

1. HERoEHS or Hanyang Erica Robot Engineering

2. NIMBro Adult Size

3. RoMeLa UCLA

4. Hephaestus Tsinghua University

Don’t forget, it’s not just football, there are also the results for quadruped challenges and RoboCup@Home and RoboCup Industrial.

Robot News

Soon a robot can do your laundry according to humanoid robotics company 1x.

Just as everyone has a smartphone, everyone will have a robot, says Bernt Øivind Børnich.

Ted Larson from Ologic dives into hardware acceleration for gen AI on the edge in robotics and I totally recommend reading the whole article. Here are two reasons why.

Benchmarks

For purposes of accelerating LLM’s or tokenized models, I HATE this benchmark of TOPS.  Most models cannot actually use the APU for acceleration due to framework constraints, and TOPS really only matters when doing vision-based AI tasks.  Tokens-per-second is the more accurate measurement.  This takes into consideration all the hardware being used, and the framework being used.  So when talking about generative AI tasks that don’t involve vision, can we PLEASE stop using TOPS as a measure of performance for applications that are not vision-based AI?   It’s driving me nuts!

Future Directions

The future of LLMs on edge AI devices looks promising due to the significant amount of research and development activity in this field, however, I think one of two things will have to happen.

1. Edge AI accelerators will move toward the acceleration of tokenized models which means access to big piles of RAM from the accelerator.

2. Large Language Model algorithms will evolve to not need matrix math, or the big weight tables that require large amounts of memory.

Agility Robotics steals the show at Fortune Brainstorm Tech, however the stories are all paywalled. Zoox CEO Aisha Evans also made the headlines making pragmatic predictions about the roll out of robotaxis.

From Brad Porter at Collaborative Robotics

Darius points out an important challenge we have in robotics today… the demos look really good, but the policies don’t generalize.

I discussed this in some depth with Prateek Joshi in this recent podcast.

XPRIZE Wildfire’s independent judging panel has selected 29 Qualified Teams from around the world to share a $750,000 milestone purse and advance to the next round of the XPRIZE Wildfire Autonomous Wildfire Response track.

Congratulations to the all the teams making it through!

Hello Robot is excited to sponsor a new competition activity PhyRC with Cornell University emPRISE Lab. PhyRC stands for Physical Robot Caregiving, with winners of Track 1 win a Kinova Gen 3 7 DOF Robot, and winners of Track 2 win a Stretch 3!

Registration Deadline: August 31, 2024
Phase 1 (Simulation) Announcement: Aug 1, 2024
Phase 1: Aug 1 - Oct 31, 2024 (3 months)
Phase 2 (Real-robot) Announcement: Nov 7, 2024 (Top-n teams released)
Phase 2 Preparation: Nov 7, 2024 - May 2025, Teams Prep for Real-robot competition (6 months)
Phase 2 Final Preparation at Venue: May 2025, Teams practice in competition arena for a couple of days (Venue TBD)
Phase 2 Competition: May 2025, Teams compete during this period and the winners are announced at the end.

Robot Events

FIRA is back! October 22-24 Woodland CA

And some more robotics conferences:

• Celebrating Women in Open-Source Robotics - Intrinsic - July 30

• Automated Road Transport Symposium San Diego July 29 - Aug 1

• Humanoid Robotics Festival Lima Peru Aug 14-15-16

• CASE 2024 Puglia Italy 28 Aug - 1 Sep 2024

• BioRob Heidelberg DE 1 Sep-Sep 4

• Commercial UAV Expo : Sep 03 - Sep 05, 2024 . Las Vegas, United States of America

• CLAWAR 2024 Kaiserslauten DE Sep 4 - Sep 6

• ACTUATE Summit : Sep 18, 2024 . San Francisco - use SVROBO for 25% off

• ICRA@40 Rotterdam NL 23 Sep - 26 Sep

• ANTS (Swarm robotics) 2024 : Oct 09 - Oct 11, 2024 . Konstanz, Germany

• IROS 2024 AbuDhabi 14 Oct - 18 Oct

• RobOpsCon 2024 - Mountain View - 15 Oct

• ROSCon Odense DK 21 Oct - 23 Oct

• CORL Munich DE 6 Nov - 9 Nov

• IEEE Future of Telepresence Pasadena CA 16 Nov - 17 Nov

• Humanoids 2024 Nancy FR 26 Nov - 28 Nov

• ProMat : Mar 17 - Mar 20, 2025 . Chicago, United States of America

• ICRA 2025 Atlanta Georgia 17 May - 23 May 2025