Alexa’s Unassuming Rise to Prominence

Alexa didn’t get a big splashy launch event, or come with a bunch of grand proclamations about being the future of anything. Instead, like a phone charger with a made-up name or a knockoff version of your favorite blush, it just kind of appeared on Amazon one day.

The Birth of the Smart Speaker Era

Ten years ago this week, on November 6th, 2014, Amazon launched the Echo and the smart speaker era. The device quickly spawned countless more, from Amazon and others. It became the emblem of a voice-first way of using technology, the “ambient computing” revolution Amazon and others came to believe would change how we do everything. Now there are Alexa devices in millions of homes, just listening and chirping away all day.

Wrestling with Alexa’s Legacy

On this episode of The Vergecast, we wrestle with what all that really adds up to. The Verge’s Jennifer Pattison Tuohy joins the show to talk about the reasons for that surprise 2014 launch, the explosive growth of the Alexa ecosystem, and the challenges Amazon and everyone else faced in trying to figure out what these smart speakers could really do. By some measures, Alexa is an undeniable hit, an utterly mainstream part of our technological lives. But for all that success, Alexa has never lived up to Amazon’s lofty goals. It’s not the ultra-powerful and ultra-versatile Star Trek computer; it’s not even a better way to shop. After all this time, it’s for music and timers. Alexa has always been for music and timers.

The Future of Alexa

There’s a big change coming for Alexa, though. We talk a lot about what the so-called “Remarkable Alexa” upgrade might mean for the virtual assistant, as Amazon shifts its underlying technology to be based on large language models and generative AI. It’s pretty clear now that Amazon’s big idea was something like the right one. Is the tech finally ready to make it real? And will Amazon ever ship the thing so we can find out? The Echo was a surprise a decade ago — maybe we’re due for another one.

Resources

If you want to know more about everything we discuss in this episode, here are some links to get you started:

Conclusion

Alexa’s unassuming rise to prominence is a testament to Amazon’s ability to innovate and adapt. As the company continues to evolve and improve its virtual assistant, it will be interesting to see how Alexa’s capabilities and impact on our daily lives continue to shape the future of technology.

FAQs

Q: What is Alexa?
A: Alexa is a virtual assistant developed by Amazon that allows users to control their smart devices, play music, and access information using voice commands.

Q: How did Alexa become so popular?
A: Alexa’s popularity can be attributed to its ease of use, affordability, and the wide range of compatible devices available.

Q: What is the “Remarkable Alexa” upgrade?
A: The “Remarkable Alexa” upgrade refers to Amazon’s plan to shift its underlying technology to be based on large language models and generative AI, which could potentially make Alexa more powerful and capable.

Q: Will the “Remarkable Alexa” upgrade make Alexa more useful?
A: It’s unclear at this point, but Amazon’s goal is to make Alexa more capable and useful, and the new technology could potentially achieve that.

By Heico Sandee, founder and CEO of Smart Robotics

Robotics and automation undoubtedly address several safety and efficiency challenges that warehouses have grappled with for years.

While efficiency and productivity are primary motivations behind automation, warehouses must prioritize the safety and well-being of those who spend most of their time around these automated systems – the employees.

Productivity and profitability mean little if warehouses are unsafe. Numbers speak for themselves as the National Safety Council reports that a worker injury costs companies an average of $38,000 in direct expenses, while indirect costs can amount up to a total of around $150,000 per accident.

Automation planning, therefore, should be centered around employee safety, striking the balance between productivity and workforce morale.

Automation can make warehouses safer for employees in contrast to traditional systems
Previously existing gaps in safety measures are being bridged by new innovations like collision-avoidance sensors, that not only handle heavy lifting but also anticipate worker movement, ensuring a safer interaction on the warehouse floor.

By taking over hazardous tasks, automation reduces the risk of injuries related to lifting, manual handling and mitigates common warehouse hazards such as slips, trips, and falls.

As a case in point, Smart Robotics’ intelligent cobot solution Smart Palletizer was implemented by an Australian milk formula manufacturer seeking to improve employee safety – a project that later received a nomination for a Safety Award for its positive impact on workplace safety.

Their previous manual stacking process involved two staff members, leading to health issues such as repetitive strain injuries and blind spots for forklifts, increasing accident risks.

The Smart Palletizer, with its compact design and easy integration into the existing processing and packaging line, allowed the manufacturer to reduce labor requirements by 50 percent, cutting costs while eliminating the safety hazards associated with manual palletizing.

Ensuring proper implementation to prevent mishaps and accidents

Thorough warehouse planning and robust safety protocols must be put in place to make sure that new systems do not inadvertently introduce new risks.

For instance, if automated systems are not properly integrated or lack adequate safety protocols, they can heighten the risk of collisions between machines and employees, leading to potential accidents or equipment malfunctions.

After all, the solution to one problem shouldn’t become another problem.

Providing warehouse employees with sufficient training to rightly interact with automated systems is a crucial step toward maintaining a safe workplace. It’s essential that there are no instances of confusion, incorrect use or unsafe interactions in the day-to-day operations.

Pedestrian safety, for instance, shouldn’t be overlooked. Warehouse staff moving between work areas, managers or supervisors conducting inspections, maintenance workers or even delivery or logistics personnel who may need to move through the warehouse on foot, all share space with forklifts, automated guided vehicles (AGVs) and other machinery, exposing them to risks of accidents.

This means that irrespective of an employee’s role or how closely they work with automated systems, everyone should receive basic training on essential protocols and safe practices on the warehouse floor.

Similarly, all employees should be familiarized with emergency procedures in the event of a system malfunction and better yet, regular maintenance and updates must not be missed to prevent such incidents in the first place.

Prioritizing mental health

Automation relieves employees of numerous repetitive and physically demanding tasks, however, such changes can also potentially introduce new stressors, including fears of job displacement, pressure to match machine efficiency, and the urgency to adapt to a highly automated environment. Therefore, change management should be considered a high-priority area.

Encouraging clear communication is one of the most fundamental tenets to be practiced to create a balanced, positive work environment. The importance of clearing the air of anxiety and addressing any concerns cannot be emphasized enough.

Regularly providing updates about changes, prompting feedback and suggestions, and creating forums for open dialogue are some of the ways in which organizations can ensure that implementations do feel enforced or one-sided.

Mental health check-ins at regular intervals, along with helpful resources such as counseling and stress management workshops, are essential during the transition phase and beyond.

These initiatives help employees navigate changes, address ensuing challenges, and promote their overall well-being. These also present opportunities for businesses to reevaluate their mental health support systems and to make warehouses less stressful places to work.

Moreover, investing time and resources in retraining and upskilling programs helps employees perceive automation as a catalyst for career advancement rather than as a threat. This contributes to greater job security and job satisfaction.

Adapting to changes is hard enough; employees shouldn’t have to stress about handling complicated systems. Automation should be designed and implemented in a user-friendly and intuitive manner, such as collaborative robots (cobots) that can work safely alongside humans.

Today, technologies are designed to directly improve safety through AI-powered sensors, predictive maintenance, and machine learning systems that proactively predict and avoid hazardous situations.

Planning for ergonomics

Workplace ergonomics has lasting implications for employees – especially so in activity-intense settings like warehouses.

Given that traditional warehousing involves substantial heavy lifting, prolonged periods of standing or walking, and repetitive movements throughout the day, it is important to understand the ergonomic challenges currently faced by employees while planning for automation.

Proposed changes must not only ensure that automated systems effectively address these issues, but they must also anticipate potential strains from new systems and work to prevent them.

For example, implementing adjustable robotic workstations and selecting cobot models designed for repetitive yet safe movements can significantly alleviate ergonomic challenges.

At its core, automation can lower the risk of hazards such as repetitive strain injuries (RSI) and discomfort resulting from repetitive body movements.

When workstations and workflows are designed consciously to fit the physical capabilities of employees, considering both their strengths and weaknesses, workers are less fatigued and can perform tasks more quickly and efficiently.

In this win-win situation, employees feel comfortable and safe in their workplaces, while organizations benefit from reduced absenteeism and a more engaged workforce.

Employee well-being

Decision makers in executive positions may not have first-hand experiences to foresee and gauge the potential impact of planned changes.

Particularly for warehouses that go from absolutely no automation to a significantly higher level of automation, it’s beneficial to include employees right in the earlier stages of automation.

Firstly, this will help identify and acknowledge employee pain points, fears and expectations regarding automation. Secondly, when employees feel involved in the process from the get-go, they are less likely to feel threatened and more inclined to embrace the collaboration.

As the primary users of automated systems on the floor, their input should be highly valued as a critical source for refining and improving these systems.

Involving employees in the easy adoption stages enables smoother transitions, eases the learning curve, and helps build a more resilient workforce.

Regular check-ins with operators regarding the user-friendliness of robotic systems and potential improvements can provide valuable insights, contributing to sustained operational excellence.

Together with upskilling opportunities, this approach prepares employees to confidently work alongside automated systems while instilling a sense of ownership and accountability over their roles in shaping a safer, more efficient workplace.

About Heico Sandee: Heico Sandee is founder and CEO of Smart Robotics, holds a PhD degree and previously acted as program manager for robotics at Eindhoven University of Technology. With more than 15 years of experience in robotics development, Heico now leads Smart Robotics in developing “intelligent, robot-independent software for flexible deployment of automated solutions”.

Despite AI Progress, Robots Remain Limited in Their Abilities

Despite stunning AI progress in recent years, robots remain stubbornly dumb and limited. The ones found in factories and warehouses typically go through precisely choreographed routines without much ability to perceive their surroundings or adapt on the fly. The few industrial robots that can see and grasp objects can only do a limited number of things with minimal dexterity due to a lack of general physical intelligence.

The Need for More Capable Robots

More generally capable robots could take on a far wider range of industrial tasks, perhaps after minimal demonstrations. Robots will also need more general abilities in order to cope with the enormous variability and messiness of human homes.

The Role of General Physical Intelligence

General excitement about AI progress has already translated into optimism about major new leaps in robotics. Elon Musk’s car company, Tesla, is developing a humanoid robot called Optimus, and Musk recently suggested that it would be widely available for $20,000 to $25,000 and capable of doing most tasks by 2040.

New Developments in Robot Learning

Previous efforts to teach robots to do challenging tasks have focused on training a single machine on a single task because learning seemed untransferable. Some recent academic work has shown that with sufficient scale and fine-tuning, learning can be transferred between different tasks and robots. A 2023 Google project called Open X-Embodiment involved sharing robot learning between 22 different robots at 21 different research labs.

Overcoming the Challenges of Robot Data

A key challenge with the strategy Physical Intelligence is pursuing is that there is not the same scale of robot data available for training as there is for large language models in the form of text. So the company has to generate its own data and come up with techniques to improve learning from a more limited dataset. To develop π0, the company combined so-called vision language models, which are trained on images as well as text, with diffusion modeling, a technique borrowed from AI image generation, to enable a more general kind of learning.

Scaling Up Learning

For robots to be able to take on any robot chore that a person asks them to do, such learning will need to be scaled up significantly. “There’s still a long way to go, but we have something that you can think of as scaffolding that illustrates things to come,” Levine says.

Conclusion

While there have been significant advancements in AI and robotics, there is still a long way to go before robots can truly perform tasks that are as complex and varied as those done by humans. However, recent developments in robot learning and the combination of different AI techniques hold promise for achieving this goal in the future.

FAQs

Q: What are the limitations of current industrial robots?

A: Current industrial robots typically go through precisely choreographed routines without much ability to perceive their surroundings or adapt on the fly. They can only do a limited number of things with minimal dexterity due to a lack of general physical intelligence.

Q: What is general physical intelligence?

A: General physical intelligence refers to the ability of a robot to perceive and adapt to its environment in a more general way, rather than being limited to a specific task or set of tasks.

Q: How will robots be able to perform tasks that are as complex and varied as those done by humans?

A: Robots will need to be able to learn and adapt in a more general way, through the combination of different AI techniques and the development of more capable robots that can take on a wider range of tasks.

Imagine having to straighten up a messy kitchen, starting with a counter littered with sauce packets. If your goal is to wipe the counter clean, you might sweep up the packets as a group. If, however, you wanted to first pick out the mustard packets before throwing the rest away, you would sort more discriminately, by sauce type. And if, among the mustards, you had a hankering for Grey Poupon, finding this specific brand would entail a more careful search.

MIT engineers have developed a method that enables robots to make similarly intuitive, task-relevant decisions.

The team’s new approach, named Clio, enables a robot to identify the parts of a scene that matter, given the tasks at hand. With Clio, a robot takes in a list of tasks described in natural language and, based on those tasks, it then determines the level of granularity required to interpret its surroundings and “remember” only the parts of a scene that are relevant.

In real experiments ranging from a cluttered cubicle to a five-story building on MIT’s campus, the team used Clio to automatically segment a scene at different levels of granularity, based on a set of tasks specified in natural-language prompts such as “move rack of magazines” and “get first aid kit.”

The team also ran Clio in real-time on a quadruped robot. As the robot explored an office building, Clio identified and mapped only those parts of the scene that related to the robot’s tasks (such as retrieving a dog toy while ignoring piles of office supplies), allowing the robot to grasp the objects of interest.

Clio is named after the Greek muse of history, for its ability to identify and remember only the elements that matter for a given task. The researchers envision that Clio would be useful in many situations and environments in which a robot would have to quickly survey and make sense of its surroundings in the context of its given task.

“Search and rescue is the motivating application for this work, but Clio can also power domestic robots and robots working on a factory floor alongside humans,” says Luca Carlone, associate professor in MIT’s Department of Aeronautics and Astronautics (AeroAstro), principal investigator in the Laboratory for Information and Decision Systems (LIDS), and director of the MIT SPARK Laboratory. “It’s really about helping the robot understand the environment and what it has to remember in order to carry out its mission.”

The team details their results in a study appearing today in the journal Robotics and Automation Letters. Carlone’s co-authors include members of the SPARK Lab: Dominic Maggio, Yun Chang, Nathan Hughes, and Lukas Schmid; and members of MIT Lincoln Laboratory: Matthew Trang, Dan Griffith, Carlyn Dougherty, and Eric Cristofalo.

Open fields

Huge advances in the fields of computer vision and natural language processing have enabled robots to identify objects in their surroundings. But until recently, robots were only able to do so in “closed-set” scenarios, where they are programmed to work in a carefully curated and controlled environment, with a finite number of objects that the robot has been pretrained to recognize.

In recent years, researchers have taken a more “open” approach to enable robots to recognize objects in more realistic settings. In the field of open-set recognition, researchers have leveraged deep-learning tools to build neural networks that can process billions of images from the internet, along with each image’s associated text (such as a friend’s Facebook picture of a dog, captioned “Meet my new puppy!”).

From millions of image-text pairs, a neural network learns from, then identifies, those segments in a scene that are characteristic of certain terms, such as a dog. A robot can then apply that neural network to spot a dog in a totally new scene.

But a challenge still remains as to how to parse a scene in a useful way that is relevant for a particular task.

“Typical methods will pick some arbitrary, fixed level of granularity for determining how to fuse segments of a scene into what you can consider as one ‘object,’” Maggio says. “However, the granularity of what you call an ‘object’ is actually related to what the robot has to do. If that granularity is fixed without considering the tasks, then the robot may end up with a map that isn’t useful for its tasks.”

Information bottleneck

With Clio, the MIT team aimed to enable robots to interpret their surroundings with a level of granularity that can be automatically tuned to the tasks at hand.

For instance, given a task of moving a stack of books to a shelf, the robot should be able to  determine that the entire stack of books is the task-relevant object. Likewise, if the task were to move only the green book from the rest of the stack, the robot should distinguish the green book as a single target object and disregard the rest of the scene — including the other books in the stack.

The team’s approach combines state-of-the-art computer vision and large language models comprising neural networks that make connections among millions of open-source images and semantic text. They also incorporate mapping tools that automatically split an image into many small segments, which can be fed into the neural network to determine if certain segments are semantically similar. The researchers then leverage an idea from classic information theory called the “information bottleneck,” which they use to compress a number of image segments in a way that picks out and stores segments that are semantically most relevant to a given task.

“For example, say there is a pile of books in the scene and my task is just to get the green book. In that case we push all this information about the scene through this bottleneck and end up with a cluster of segments that represent the green book,” Maggio explains. “All the other segments that are not relevant just get grouped in a cluster which we can simply remove. And we’re left with an object at the right granularity that is needed to support my task.”

The researchers demonstrated Clio in different real-world environments.

“What we thought would be a really no-nonsense experiment would be to run Clio in my apartment, where I didn’t do any cleaning beforehand,” Maggio says.

The team drew up a list of natural-language tasks, such as “move pile of clothes” and then applied Clio to images of Maggio’s cluttered apartment. In these cases, Clio was able to quickly segment scenes of the apartment and feed the segments through the Information Bottleneck algorithm to identify those segments that made up the pile of clothes.

They also ran Clio on Boston Dynamic’s quadruped robot, Spot. They gave the robot a list of tasks to complete, and as the robot explored and mapped the inside of an office building, Clio ran in real-time on an on-board computer mounted to Spot, to pick out segments in the mapped scenes that visually relate to the given task. The method generated an overlaying map showing just the target objects, which the robot then used to approach the identified objects and physically complete the task.

“Running Clio in real-time was a big accomplishment for the team,” Maggio says. “A lot of prior work can take several hours to run.”

Going forward, the team plans to adapt Clio to be able to handle higher-level tasks and build upon recent advances in photorealistic visual scene representations.

“We’re still giving Clio tasks that are somewhat specific, like ‘find deck of cards,’” Maggio says. “For search and rescue, you need to give it more high-level tasks, like ‘find survivors,’ or ‘get power back on.’ So, we want to get to a more human-level understanding of how to accomplish more complex tasks.”

This research was supported, in part, by the U.S. National Science Foundation, the Swiss National Science Foundation, MIT Lincoln Laboratory, the U.S. Office of Naval Research, and the U.S. Army Research Lab Distributed and Collaborative Intelligent Systems and Technology Collaborative Research Alliance.

Google Photos Gets New Video Editing Features

Google Photos continues to get new and improved features on a regular basis, and one of the most recent Android updates has focused on video editing. Even if you don’t have the latest Pixel 9 phone (which is required if you want to try out the weird new Reimagine tool), you can now speed up, slow down, and enhance your clips with a few taps as well as trim them down more easily.

Trim Tool

First up, we have what Google describes as “improved controls” for cutting out extraneous footage at the start and end of your clips — though, to my eyes, there’s not a huge amount that’s different here compared to the previous version of the trim tool.

The handles at each end of the clip are a little bigger and thicker, making them easier to hit with a finger press. You also get a timestamp shown onscreen as you drag those handles around, so overall, the edits are a little easier to apply.

Using the Trim Tool

• Drag the left-hand handle to change where the video starts.
• Drag the right-hand handle to change the video’s end point.
• Drag the white bar between the two handles to move around the clip.
• Tap the play button at any point to check your new footage.
• Choose Save copy to confirm your changes and save a separate clip.

Auto-Enhance Tool

Google Photos now has a new auto-enhance feature you can access, which analyzes your clip and then applies its own choice of color enhancements, while stabilizing the video at the same time.

Using the Auto-Enhance Tool

• Tap Video then Enhance to apply the automatic enhancements.
• Tap the play button to see how the updated footage looks.
• Tap Enhance again to see the difference with and without the tweaks.
• Choose Save copy to save the enhanced video as a separate file.

Speed Tool

The new speed tool in Google Photos gives you more control over the pace of your videos for speeding up and slowing down the action. The effect can be applied to a specific section of your clip or all of it.

Using the Speed Tool

• Tap Video and then Speed to bring up the editor.
• Use the bars on the timeline to indicate where you want the effect to start and stop.
• Choose a playback speed under the timeline: from 1/4 speed to 4x the speed.
• Tap the play button to see how the video now looks.
• Choose Done, then Save copy when you’re happy with the results, to save a separate video file.

Video Presets

Unlike the tools above, which are exclusive to Android, the AI-powered video presets are available in Google Photos for both Android and iOS. Or at least, they will be eventually — though they were announced in September, as of this writing, I haven’t yet seen the presets in the Google Photos apps on either platform.

Using the Video Presets

When they do appear, the Presets button will appear between Video and Crop in the options at the bottom of the interface. Select it, and you’ll see a choice of edits you can apply with a tap: Basic cut, Slow-mo, Zoom, and Track. These will be applied as the Google Photos AI sees fit based on the video content.

Conclusion

Google Photos is continually updating its features to make it easier for users to edit and enhance their videos. The new trim, auto-enhance, and speed tools are a great start, and the AI-powered video presets are sure to be a hit once they’re available.

FAQs

Q: Is the Reimagine tool available on all Android phones?
A: No, the Reimagine tool is only available on the latest Pixel 9 phone.

Q: Can I use the new video editing features on my iOS device?
A: Yes, the AI-powered video presets will be available on both Android and iOS devices, although the other new features are exclusive to Android.

Q: Can I use the trim, auto-enhance, and speed tools on my iPhone?
A: No, these tools are exclusive to Android devices.

Q: Can I apply the video presets to a specific section of my clip?
A: Yes, you can apply the video presets to a specific section of your clip, or to the entire clip.