Sunday, October 17, 2021

Sad News

I recently learned that Lee Gomes, the journalist who published the most detailed and accurate report on issues around uBeam, had died. I came to know Lee after I left uBeam and talked with him often, though in the last couple of years clearly less than I should have. Lee was a great tech journalist, determined to get to the truths of what he was looking into, always being sure to get good sources and triple check everything. He was also just a good person, and I had many calls with him covering topics like US and UK politics - he seemed to especially enjoy talking with me about Brexit.

There's a detailed obituary on his life and career here.

Sad to hear this, and a great loss.



uBeam and MEMS Issue

 A recent post by a former uBeam engineer made an interesting comment:


"The biggest annoyance at uBeam was if the demo units were running when you took a call. The other side was always like "what the HELL is going on over there?" "Oh yeah sorry I'll go in the other room." The ultrasound down-mixes in the MEMS mic."

I'd posted before on the potential effects of ultrasound on gyros and MEMS, this is the first I've heard of uBeam affecting phones this way. I'm sure this information was relayed to customers and investors for them to take into account...



Monday, July 12, 2021

Dead Startup Toys

For all of you that wish you could own one of the cool dead startup gadgets that gets written about here, then Dead Startup Toys will do it for you - well, kinda. 

Toy replicas of failed companies including Juicero and Theranos, though selling out fast as the Coolest Cooler and Jibo are already gone. There's a small "bio" of each product as you click through. Nothing from uBeam or Energous there though, sad to say. Will June or Sunflower make it there? I'm interested to see their next toys.

The group making these is really interesting - MSCHF - I think this might be my favorite thing they've done:

In February 2021 the group purchased a Boston Dynamics robot dog, and mounted a paintball gun on it. The robot was used in a live performance that allowed users of the MSCHF app to control the robot and its paintball gun. After MSCHF publicly criticized the potential use of robotic dogs by police forces, Boston Dynamics released a statement criticizing the use of the robot in an artwork


Friday, May 7, 2021

Travel - Then vs Now

I'm fortunate enough to be fully vaccinated, so a couple weeks ago I took my first plane flight in over a year. For most of the prior 15 years or so I've flown around 50 to 100,000 miles a year, so a moderate flyer (given 100,000 miles flown is around 200 hours in the air, or over a solid week, it's still pretty significant). I went and visited friends in the NE of the US, and damn do I miss socializing and just hanging out with people. Airports and planes are as awful as you remember, but they get you to places and people that are worth it all. Before I left, I looked at my phone for a map of where I'd been in the previous 6 months:

You might be able to spot where I go vaccinated. King County, WA, where I live, you couldn't find an appointment for love nor money - but drive out a couple hours to rural eastern WA and you had your pick that day. It's also a really scenic drive. And here is the equivalent time period two years prior:

I think my personal carbon footprint was a little larger in 2018 than 2020. Despite the pain of flying, I'm still looking forward to getting out there again, and as soon as the UK drops the test/quarantine on arrival, I'll be heading back there. Since I moved to the US I'd been back in the UK once or twice a year and saw family, 2020 was the first year I hadn't and so it'll likely be a 2 year gap. Fortunately, despite being a basket case in most other regards, the UK did a great job with vaccinations (thank you NHS, yay socialized healthcare), and my mother got her first vaccine in mid January, which for those of you with elderly relatives will know the relief. 

For those in the US who wonder about the value of a good single healthcare system, the NHS in the UK prioritized similarly (primarily age) but literally called up each person and sent letters saying "It's your turn in 2 weeks, here's your appointment" and proceeded that way. The UK also took the route of making the gap between the two shots up to 12 weeks, not the 3/4 in the US, to maximise the number of people with one shot. From people I know the gap seems closer to 9 weeks in practice.

For those in the UK wondering the US in comparison (in the way back of a few weeks ago when there were more supply restrictions) you were pretty much on your own - vaccines were spread between some pharmacies, some hospital vaccination sites, some mass vaccination sites, and you needed an appointment, which was all done through a mish-mash of systems and you needed to be somewhat tech savvy, have time on your hands, and be willing to move quickly which basically was the exact opposite of what many key groups such as the elderly or front line workers with multiple jobs had. And everyone seems to know someone who skipped the line because they knew someone who knew someone, or worked for the right company (I knew work-from-home accountants at Amazon, with no morbidities, who got it in February because... well they were Amazon. And I knew people with health issues who had to wait.), or if you loitered around pharmacies at closing time and got leftover doses. Now in the US we've reached the point where any adult can get it via walkin - amazing the effect of competency at the top - there are more doses than willing people, thank you anti-vaxxers for making it harder.

Maybe I'll post my equivalent travel again in a year. I hope it's closer to that second picture than the first, because that means our world has returned at least somewhat back to what it was before.

Tuesday, May 4, 2021

My Spy in uBeam

Now that uBeam is dead, I can reveal a secret that I have kept close to my chest for years. You see, for some time after I left the company I had a spy in uBeam, someone who quietly passed information to me about who they were interviewing, companies they visited and were discussing deals with. Management eventually realized this information was being leaked, and despite some significant efforts to root them out the spy evaded the devious traps that were set for them. 

Regardless of the danger to them I now have to let everyone know - that spy was...


Meredith Perry


Yep. It was the co-founder and CEO herself. You see I was still a first connection to Perry on LinkedIn after I had left the company, and she was quite diligent about connecting with every person that was interviewed, every company executive, manager, investor, or similar she spoke to. I didn't even have to actively go looking, the connections just kept appearing in my feed, as LinkedIn especially at that time just shoved them in your face. 

Knowing the industry, I had a pretty good idea of what the conversations might be about or the positions the person was interviewing for. Journalists would regularly call me at that time with questions about the company, and I would mention companies I thought they might be speaking to (avoiding any that I might have learned about while employed there). I guess the journalists called them every now and then and sometimes the people in question would ask Perry why they were contacted, thus revealing the presence of a spy.

Now at this point I have to apologize to a number of the staff who were still there at the time, because Perry immediately assumed one or more of them was leaking (to be clear, no-one at uBeam I knew socially ever told me company proprietary information, and I didn't ask). They were repeatedly warned, grilled, and placed under suspicion to a point I later learned might even have involved Perry having IT give her the access to read all the emails of the staff each day.

This does not even cover the stuff Perry would just blurt out on calls with people in the industry I knew - she'd call up people and basically just ask them to work for uBeam, talking about what they were doing, and what deals were in the works, all without an NDA. I'd then get an email about it or hear at the next conference.

So the moral of the story? I have no idea. I just still find it all funny to this day.


Monday, May 3, 2021

It's an ex-uBeam...


Just shy of its 10th birthday and with between $40 and $48 million of investment (estimated), it appears uBeam (recently Sonic Energy) has shuffled off its mortal coil. While some may claim it's simply pining for the fjords or merely stunned, I've heard that a few weeks ago the last remaining employees were told they were terminated effective immediately and the doors were closed. Whether the company will actually be killed, or carry on in zombie form as an asset holding entity remains to be seen, but reportedly it will be OurCrowd, the crowdfunding group who were significant investors in the later rounds, that take possession. 

I'm sure co-Founder and former CEO Perry will claim that the company was taken from her too soon and would have succeeded if only under her watch, or it was those pesky engineers just not implementing her vision, but to use her own phrase  "Our industry is binary, your tech works, or it doesn't,". In this case, it doesn't. 

So what did they have? From descriptions given to me by those who saw it in operation like prospective customers, there were moderately bulky panels that transferred power on the milliWatt level (a phone charges at ~1 Watt or more) at around a meter, with limited steering and tracking. I never got to hear any efficiency numbers, but I'd guess single digit % at best. Under those conditions the question to ask is "How long will a battery the size of the receiver last with the same milliWatt draw?" and as always "If it's not moving will a wire be simpler, cheaper, and more reliable?"

Where this might work is a low power demand system like remote sensors, you can get close to and in line of sight, that don't move much, where batteries aren't allowed, and you can't run a wire to. Pretty niche market.

Recent patent applications (not grants, here and here) show a system made up of hexagonal modules, each with a speaker like cone attached. I'd be interested in seeing a side by side comparison with the Murata S40 devices as it is a very similar principle, I suspect they would be comparable, perhaps slightly more narrowband, and a bit more expensive. 

This doesn't even begin to address the question of safety - if it's at 145 dB it certainly exceeds limits in most countries and it looks like also in the US as of today. Further, the transducers are also quite wide, and unless they send most energy straightforward (which would limit steering) then they will generate grating lobes when used as an array and send energy at undesired directions. Perhaps now it's defunct uBeam can release the claimed studies by independent third parties that "proves" it is safe, but bizarrely were never made public.

It's shame that no-one sensible told the board in summer 2015, after "only" about $8 million had been spent, that leadership was taking the company in the wrong direction, increasingly "exaggerated claims" on performance were being made, and that a pivot was needed. But here we are - end of an era. 

Saturday, May 1, 2021

Quite the 18 Months

My last post was December 2019, putting this at nearly 18 months since I last wrote anything here. It's not for lack of interest, but simply I was massively overloaded with work, commitments, and personal issues that meant something had to go, and so it was the blogging. Things are somewhat more stable now, so I'm intending on getting back to this on a semi-regular basis. Interestingly my first post on this blog was 15th April 2016, so just over 5 years now since I started it. Long enough that my NDA with uBeam has now expired - though don't expect any changes in what I write, things traditionally and reasonably covered by an NDA are not things I will talk about even on uBeam.

So what have I been up to? Well, I quit my full time job at the beginning of 2020 to concentrate on a company I co-founded with the other VP Engineering from uBeam, Sean Taffler. The company is Acoustiic, and our goal is to enable the next generation of non-invasive cancer surgery using High Intensity Focused Ultrasound (HIFU). 

This is a well known technique, currently growing in applications, with an increasing number of FDA approvals and insurance reimbursements - if you want to read more there's a great resource at the Focused Ultrasound Foundation, a non-profit dedicated to enabling better medical outcomes from ultrasound. What we aim to offer is the next generation of systems for HIFU - more precise, more controllable, faster - using advanced manufacturing techniques, custom electronics, and leveraging the compute power available today. Sean did an interview with the FUS Foundation on FDA approvals recently, you can read it here.

The other thing that was eating up my time is that starting in 2020 I became both the President of the IEEE Society for Ultrasonics, Ferroelectrics, and Frequency Control (UFFC) and the General Co-Chair for our largest conference. Each is a fairly hectic job under normal conditions, but add in starting a new company and a pandemic and everything goes to hell - shifting a ~2000 person conference from in-person to all online at short notice, while negotiating with the original venue not to lose a large deposit, was certainly a novel experience. It was a lot of work, but also gave us the opportunity to pull forward online presentation options years faster than would be possible otherwise. For those interested in some of the things we did, there are some newsletter articles I wrote for UFFC that can be found here (May 20, August 20, November 20, April 20). I also did an article on SBIR (Small Business Innovation and Research) grants for the IEEE, it's here and I will also post here as a blog post now it's been up a while.

There are still a few uBeam stories to be told, and I'll get to them over time, but I'll likely be transitioning to talking about other startup and fundraising topics since that's the world I'm living in now. So, let's see if I can do a little better than 18 months for my next post...

Friday, April 30, 2021

SBIR – Non Dilutive Funding for Your Company

This is a reblog of an interview I did last year for IEEE-USA, the original can be found here.

What if I told you there is a funding opportunity in the US that is restricted to small companies, focuses on technology, and takes no equity (non-dilutive), debt stake (grant or contract, not a loan), or claim on Intellectual Property (IP) for what may be a seven figure investment? The SBIR – Small Business Innovation Research – program has been operating since 1982, providing funding totaling over $2 billion per year, and giving early starting capital to companies that grow from there including the likes of Symantec, Qualcomm, Intuitive Surgical, and Illumina. 

History

The SBIR program came into law in 1982, founded by Roland Tibbetts, whose goal was to fund innovation, supporting American companies and giving them the support to take ideas to a stage where other private investment can take over. It was last renewed as part of the 2017 National Defense Authorization Act and is currently authorized to September 2022.

All federal agencies with a large R&D budget are required to set aside part of that budget (3.2%) exclusively for small businesses, with 500 or fewer employees, that are majority owned by Americans and permanent residents. There are 11 agencies that administer SBIRs - the Department of Defense (DoD) and National Institutes of Health (NIH) are by far the largest and account for nearly 70% of the ~$2.5 billion of funding, then Department of Energy (DoE), National Science Foundation (NSF), and NASA provide under 10% each, and the remainder is split between the other 6 agencies. Application processes vary between agencies, but there are common factors between them.

The Small Business Technology Transfer (STTR) is a sister program to SBIR. It is similar but requires participation of an academic partner, is restricted to a smaller number of agencies, and with a smaller set-aside. Aside from these differences, the broad strokes of SBIR also apply to STTR.

SBIR Phases

Most SBIRs start with an application for a Phase I – often a proof of concept – that is expected to take between 6 and 12 months, with funding for $100,000 to $400,000 (depending upon agency and project area). Applications usually take the form of a 6 to 12 page technical proposal, a project plan, a commercialization plan, a single page budget, personnel biographies, letters of support, and other relevant information. The proposal is reviewed by a committee, and typically awardees are notified and funded within 4 to 6 months. By the end of Phase I, the company has ideally demonstrated a proof of concept, and its competency, to take the project further.

Phase II is a more substantial effort with greater budgets – up to around 3 years and $2,000,000 (again, depending on agency and topic). Usually an agency invites a company to apply for a Phase II after a successful Phase I – there are some Fast-track and “Direct to Phase II” options but are unusual. The application is usually much more detailed, and budget and timelines come under more intense scrutiny, and so is recommended for more seasoned applicants. By the end of Phase II, a successful project is usually expected to have something between a working prototype and a commercial product. 

Phase III is the commercialization phase and receives no specific SBIR funds, though some agencies do have matching and follow on funding. Agencies with very specific needs often try to ensure valuable technologies make it to market, such as the DoD connecting a small business with primes like Boeing.

With a few months between Phase I and II, it can take as much as 5 years to get from application to the end of a successful Phase II – these are not small projects or commitments!

Application Process

Companies applying for SBIRs need to follow a few administrative basics – the first is setting up a for-profit company (e.g. an LLC, C-corp etc) in the US, with >50% ownership by US citizens or permanent residents, and obtaining an Employer Identification Number (EIN) from the IRS. The company should then apply to Dun and Bradstreet for a DUNS – a unique identifying number. (They may try to sell you other services. Just keep telling them you need it for a federal grant application and you get it for free). The next step is registering with the System for Award Management (SAM) to apply for government contracts. As you can see, all this takes some time and effort, and can take a couple of weeks to complete. The SBIR.gov site has a walkthrough to help here.

Each agency has a specific application site. NSF uses Fastlane, the NIH uses eRA Commons, the DoD uses DSIP etc. Each site has its own registration requirements for the company and potentially the key investigation staff, that also may take a few days to setup. I usually tell applying companies to allow a month prior to submission to get this process completely done the first time. Subsequent applications are then much faster.

Each agency like the DoD, NIH, NSF, DoE, and NASA maintain their own SBIR websites, while there is also an overall SBIR site, where more details on submission can be found.

Every proposal has key staff and a Principal Investigator (PI) who must work at least 50% of their time at the company. Proposed budgets may cover items such as salary for key staff, consultants, equipment, and materials, justifiable travel, as well as reasonable overhead costs. The majority of the budget must be spent within the company (not outsourced), and any outsourced work should be done within the USA.

Agency Specifics

As mentioned, each agency administers the SBIR in its own way. The DoD has very targeted needs and this is reflected in its methodology. Twice a year the DoD SBIR topics are opened up, and each branch has a number of targeted topics (For example, the Navy may call for development of a more compact and higher performing SONAR buoy. The application deadline is about 2 months after the topic release, after which the topic is gone.

The NSF has a different approach, with 4 deadlines a year, where a submission on any topic within its remit is encouraged. Companies can pre-submit a short description to the NSF, and receive feedback on whether an application is warranted or not.

The NIH has a mix, with three submission dates a year for most applications, which similar to the NSF can be anything within its broad remit. Applicants are usually encouraged to speak to the SBIR manager in that area to gauge general response, and as a courtesy to let the NIH staff for review. The NIH also has calls on specific topics that run for a limited period, for example one to two years, on areas of particular need. Dates on some topics may vary slightly.

How Agencies Benefit

What do the agencies get for this money? First of all, they don’t get equity in your company, they don’t get paid back, they don’t own any product or IP that is created, and provided you label everything sent to them in reports they don’t own data rights either.

They do get research and products in areas critical to them. The products may not sell, or the research may show an approach is not viable, but it’s understood that there is risk involved. You owe them the reports, milestones, and deliverables listed in the proposal. Even then if the ongoing research shows a pivot is needed, every monitor I have been involved with has been very reasonable about changing course.

In most cases the project monitors are have spent their career in these fields and care deeply about curing disease, advancing technology areas, or helping those in the military get the best equipment and tools, or whatever their area is. They aren’t in this for personal reward, and they want you to commercially succeed because it means more and better products in their key areas. In my personal experience, almost every government agency representative is an exceptional advocate for their field, and works hard to get positive outcomes for the company, agency, and the country.

The other side of this is that the funding for SBIRs comes from an overall pot, not from an individual group’s budget. Getting awards in their area effectively increases their departmental budget, and your success benefits them in multiple ways. They *want* you to be the next Qualcomm.

This is one of the few times you will have a counterparty who doesn’t care solely about the bottom line.

The Downsides

So far it seems easy – write a few pages, join some websites, and you get up to $2.4 million. These awards are competitive - there are many others applying, and your application is only one of many other worthy contenders. NIH Phase I submission success rate averages around 15 to 20%. Phase II averages around 30 to 40%, and there is no guarantee you will be in that pool. An excellent idea with a top notch team and a well written proposal will put you above that average, but you are unlikely to be there on your first application. When I work with companies writing SBIRs, I advise them to consider them for the long term, and that several attempts will be needed before they succeed. 

As with anything new, the first proposal you write will take some time. Just the administrative preparation can take a few weeks, as you work with various websites and develop the company boilerplate sections, key member bios, and learn budgeting methods.  You’ll also need to write compelling research and commercialization plans. Companies should expect to invest about 6 man weeks into writing their first SBIR proposal (apart from the administrative setup), 2 weeks into the next, and a week every one after that. As the business minded among you are calculating, it does not make sense to have one of your most important staff members spend 6 weeks for a 20% chance of $150k – but if you are in for the long haul and assume this is going to continue over years, then it quickly shifts to being worthwhile. Over time, you will learn not only how to write SBIRs but when to apply and which ones not to. Ultimately, you can get to the point where a few days' work on a well targeted application gets you a >50% chance of success.

After receiving the award, you are subject to Federal Acquisition Regulations (FAR), which require careful financial tracking and reporting and limits how you can spend your funding. It is not a pot of money that can be used any way you want, but a well-run company with good accounting should have no problems meeting FAR.

While the federal government will always pay, sometimes it takes time for federal payments to arrive, or for contracts to be renewed. This can leave periods where your company is left without income but still has employees. While agencies try to minimize this, it is a concern. If possible, try to maintain sufficient revenue or cash in your company to pay staff during these uncertain times.

The Upsides

There are several upsides beyond “up to $2.4 million of non-dilutive funding”. The agency is a customer, giving you input and feedback as you develop, and may ultimately be a source of revenue. This third party endorsement proves to others your idea and team have merit, and if you are raising money from other sources, helps validate you.

Even if not initially successful, the process of creating the research and commercialization plans often focuses a company on where they are and where they need to go I’ve often found that this is the first time a company has laid out a formal plan for their technology and business. Further, the reviews returned can be eye opening as to what other experts feel are the strengths and weaknesses of your approach, and can direct future improvements. Of course, there’s always the chance that “Reviewer 2” is simply an idiot who doesn’t appreciate your genius, but that’s been grant applications throughout history!

One often forgotten benefit of SBIRs is that because they are competitive, if you have won one in a particular field, you have now met the competitive requirement in any such federal contract, so it can be a rapid way to beat your competition to the punch.

Writing a Successful Application

Beyond having an excellent idea and a great team that meets the need of a federal agency, how do you write a good application? When writing my own, the first thing I think is that I’m not writing a report, I’m telling a story. I’m leading a reviewer on a journey so by the end they know this proposal needs to be funded. I start with a title that combines an obvious and important problem and solution, and catches the eye and makes them want to read the first paragraph. The first paragraph conveys the core of the proposal over and makes them want to read the rest of the first page. Then I have the rest of the first page to make it obvious that this is the solution that’s needed. That first page is critical. The rest is only showing that you have the background and team to make this work, and not giving them a reason to doubt you. Have people you trust outside your company read your first page, and find out what excites them and why, and build that feedback into the next draft.

Keep the proposal simple and targeted. A straightforward application that picks one clear goal is easier to explain but more importantly does not give “targets” for reviewers to pick on and make negative comments. Remember that when you are up against other good applications, one small criticism may cause the agency review panel to pick another awardee over you.

Lastly, remember these are reviewed by experts who are volunteers. They have day jobs and this is usually done at midnight after a long day working. Make it easy for them to read. Use large font, short paragraphs, lots of figures and bullet point lists so they don’t have to work to understand your genius. Remember, nobody likes opening a document to see 10 pages of nine point font!

Summary

The SBIR program is a great way of accessing millions of dollars of non dilutive funding, feedback from experts, and gaining a prospective customer. It involves a lot of upfront work and discipline, but a consistent approach can yield a significant benefit to your company.

Thursday, April 29, 2021

Talking to Journalists

Something I learned in the early days of writing this blog is "how to talk to journalists", and so I thought I'd do a post on it partly to remind me in future when I forget.

First thing - you do not have to talk to them. If you are uncomfortable, then don't. For the most part, they need you more than you need them. If you don't speak to them, you are exactly where you were prior and you carry on with your day to day life - but they don't have a story (or a weaker one) and those things they need. So - again - if you don't want to, just don't talk.

Journalists are not your friend even if they are friendly. They have a job to do, and you're a source of information for them in getting that done. It doesn't mean they are hostile or out to get you, but don't speak to them like they're a friend. 

Remember that like in any profession there are good ones and bad ones. Most I've spoken to were professional, checked all pieces of information given, went to multiple sources, and tried to make sure they were writing a good story. Some clearly didn't care. Before you talk to them go look up their previous articles, the publications they work for, and see if they are well written pieces and reputable publications. If you don't like how they write, don't speak to them or do so very very carefully.

Choose your words with care, they might end up on the front page of a national paper. Pause, think, and imagine what someone you don't know will read into your words, including future employers.

Email is also a good approach to communication - you have time to think about your answers and edit, and in the end there is no question of what you did or did not say. Journalists seem to prefer voice communication though, at least in initial discussions.

Also remember that anything you say can and may be used in their story with attribution, even if you say "this is on deep background" and then tell them something. They have to agree to that before you tell them something, otherwise they are under no obligation to do so. So, if you want to do the "off the record" thing, get that clear agreement first, (possibly in an email), and also make sure you understand what the terms actually mean. Taking from Associated Press definitions:

On the record (Default): The information can be used with no caveats, quoting the source by name. 

Off the record: The information cannot be used for publication.

Background: The information can be published but only under conditions negotiated with the source. Generally, the sources do not want their names published but will agree to a description of their position. AP reporters should object vigorously when a source wants to brief a group of reporters on background and try to persuade the source to put the briefing on the record. These background briefings have become routine in many venues, especially with government officials.

Deep background: The information can be used but without attribution. The source does not want to be identified in any way, even on condition of anonymity.

In general, information obtained under any of these circumstances can be pursued with other sources to be placed on the record.

I've sometimes taken the approach of "I'll speak to you on Background, where you can describe me as 'a person familiar with the company'. If you want a quote attributed to me, I'll consider and send you it in email" because I want control on what my name is attached to. I've found most good journalists agree to this or something similar. Ones who are looking more for hits than a solid story drop you quickly - I'm not sure if that's because this makes more work for them, or because it signals they are dealing with someone who knows the game.

If you think a Non Disclosure Agreement applies to anything you are talking about, then you really, really want to talk to a lawyer first. If the lawyer says it does apply, then you do not talk unless you want to risk a lawsuit. If you still think you need to talk to them, then ask your lawyer for advice on how to do so. If in doubt, just don't. And if you can't afford a lawyer for an hour (~$500) then you can't afford a lawsuit and need to stay quiet. Similar rules apply if you are speaking regarding a person with deep pockets or access to a lot of someone else's money, even if there is no NDA, as they can still cause a lot of pain legally. 

If you've signed a government secrets act and this applies, then be prepared to go to jail for a very long time. Seriously, just don't risk this one.