Four brand-new startups from Japan, fresh off several months of intensive pitch-coaching from Silicon Valley entrepreneurial experts, took the stage on Thursday night to show off both their cutting-edge technology and everything they’ve learned from weeks of training and preparation.
The companies featured were:
- Rapyuta Robotics – a cloud robotics platform that simplifies robotics application development, enabling people with even minimal experience with robotic hardware and software as an integral part of their business ecosystem
- WFC (Wonder Future Corporation) – an innovative manufacturing company light who have created an unbreakable integral 3D resin touch panel, and electromagnetic IH (Induction Heating) packaging technology and equipment for non-heat resistant materials like PET, paper, and cloth.
- JEPlan – a circular economy company who recycle used clothing and PET bottles into polyester pellets (raw material of fashion fabric)–the same quality as the virgin PET pellet made from oil, and permanently recyclable.
- Synthetic Gestalt – in stealth mode, public info coming soon!
In keeping with the theme of emerging and future-thinking technology, the evening also featured a panel discussion on current trends in enterprise AI, with three industry experts giving their ground-floor take on the trends:
- Angelo Del Priore, Partner at HP Tech Ventures, HP’s venture arm seeking world-defining tech being built right now
- Ryohei Fujimaki, Founder and CEO of dotData, and end-to-end data science automation platform
- Manju Devadas, CEO of Pluto7, a Google Cloud AI/ML partner
A few highlights from the panel are below:
- The machines aren’t coming to take your jobs (not yet, anyway): AI is still a limited process, with most of the potential front-loaded in its business value instead of data science. “People are crazy about data scientists–it’s a ‘sexy’ job,” said Ryohei. In reality, data science is still supported by several different schools and industries, including visualization and enterprise.
- The new computer age has arrived…and it’s still arriving. “Since the 1980s, the power of connection has been driving innovation,” said Manju. It’s only been a few short years since connected devices changed from a feature to an expectation, and we’re starting to hand more and more decision-making and outcome control to the machines as we build them. The definition of AI as it currently stands vacillates between complete control and simply influencing the decisions made by machines. “What you have to realize is that sometimes getting something even 1% more accurate is still a huge step,” Ryohei said.
- AI is still working on extracting the “fun” from “fundamentals.” We’re not building Ava just yet–the panelists agreed that AI engineers are currently focused on facilitating better-quality decisions with higher accuracy. We’re seeing tech advancements already that we didn’t have 2 or 3 years ago–but, as Ryohei pointed out, “everything in AI is still manual, because we’re still adding in and processing data.” The accuracy is increasing, but it’s far from ideal–even as ROI is increasing in particular areas, the manual aspect of AI is a very particular challenge: it’s not sustainable.
- Mistakes are the hardest thing to learn from, especially for machines: “Examining problems retrospectively is important in human learning,” Ryohei said, “and for machines, I think this will become more important than accuracy in the years to come.” Truly intelligent machines won’t get locked into permanent decision paths–but in order to do so, they have to understand what a mistake is to begin with.
- AI implementation is a C-level power move. Executive decisions will be fundamental to AI implementation, since company decision making is top-down. “Companies struggle to innovate and make the new models fit the old ones. Is it a project that’s going to be sustainable, or is it a pet project? These are the questions that slow down adoption,” Angelo said. Even forward-thinking companies still struggle with the challenges presented by AI adoption.
- Privacy challenges continue to confound (and the definition of privacy itself is changing). “The challenge is not on the tech side,” Ryohey said–the philosophical side of privacy is more challenging. AI thrives on data, and has found a particular foothold in data-driven services industries like insurance and healthcare–but how much is too much? Where does process stop and human right begin? The train has already left the station, as far as machine intelligence is concerned–as Manju put it, “once you create a model that’s faster and safer, there will be no going back to the flawed models of the past.” The challenge for data scientists is to keep the human side in sight as they build the machines that serve us in the years to come.
Some comments have been edited for clarity and/or brevity.
About our international and immersion programs: Silicon Valley Forum creates unique, high-quality immersion programs, startup workshops, business luncheons, pitch nights, and more for corporations, accelerators, government and universities worldwide. Each program can be tailored to your delegation’s specific needs, find past Silicon Valley immersion programs here and download our program brochures to learn more about what we can do for you!
Director of Communications