Wednesday, February 8. 2017
Facebook is closing hundreds of its Oculus VR pop-ups in Best Buys after some stores went days without a single demo
Well, that was quick - Broadstuff predictions for Tech 2017, No. 11, Dec 31 2016:
Hate to say "We told ya", but.....
Friday, February 3. 2017
After our systems predicted Trump's win, we were asked a number of times about the impact of Fake News (and Bots, Russian Hacking etc - we will cover those in separate posts) and here is a summary of some of the useful research we looked at:
Stanford/ NYU Research
Firstly, research by Hunt Allcott of NYU and Matthew Gentzkow of Stanford, published by Stanford University looked at the sources and takeup of Fake News. They defined “fake news” as "news stories that have no factual basis but are presented as facts". By news stories they meant stories that originated in social media or the news media, i.e. excluded false statements originated by political candidates or major political figures. They also excluded websites well-known to be satirical, such as the Onion.
Firstly, they found that in the US elections, people mainly got their news by from sources other than websites and social media (see pie chart below, left). But online media (websites and social media) was where most Fake News was disseminated. They also looked at how Fake News was disseminated on the online media (below, right) and the majority was transmited via social media with a significant minority going direct (to websites or their feeds) or finding it in search results, This contrasts hugely with how top news was disseminated, mainly via older channels but online the major source was via direct access and then search.
They also looked at how people reacted to Fake News, ve Mainstream media news, and also inserted Placebo news (stories they made up) to test reactions. The chart below shows how people reacted:
The Figure presents the share of headlines that survey respondents that recall seeing (blue bar) vs. recall seeing and also believing (red bar). They averaged responses across all the headlines within four categories of headlines they presented - "Big" true stories; Smaller true stories; Fake stories and Placebo stories that they had made up headlines for. In short they found that 15 percent of people reported seeing the Fake stories, and 8 percent reported seeing and believing them (about 55%). But the chart also shows a number of other interesting tendencies:
The last test they did was to model what impact Fake News would have had to make to shift opinion in the most closely fought wards to ensure the Democrats won. For Clinton to have won the election, Trump’s margin of victory would have to decrease by ~ 0.51% of the voting age population, which would have shifted Michigan, Pennsylvania, and Wisconsin into Clinton wins and deliver the Electoral College. Thus, the core question was whether fake news could have increased Trump’s margin of error by more than 0.51 percent of the voting age population. The table below summarise the outcome of their model. In summary, the column on the far right looks at how many times more effective the Fake News would have had to be compared to TV advertising to have had to have shifted the vote. For example, on line 1 a Fake News story as it performed in reality was would have had to be 37 time more effective to shift opinion. If recall was 7% of all stories, it would have had to be 27 times more effective. Line 8 sows that if Fake News shares were 20x greater it would still have to have been 13 times more effective
Their overall conclusion was that Fake News was very unlikely to have had a major effect:
Another study was done by IPSOS for Buzzfeed on the impact of Fake News on Facebook, as Facebook had by far the largest reach of any social network for Fake News (see study here) and conclusions were in line with the above work:
IPSOS Online survey of 1,007 American adults
Percentage of consumed news in the past the month by channel showed
Print, TV and Twitter was relatively more trusted than Facebook
Far lower trust of news on Facebook all or most of the time
However, other research by IPSOS suggests that trust is not the same as belief — Another poll by Ipsos/BuzzFeed News foundon average about 75% of American adults believed fake news headlines about the election when they recalled seeing them. This contradicts the Stanford finding of c 55%, but as their model showed, even that belief level would not have changed the election outcome
In short, both studies show a minority of news was received from the online world, and it was by and large not widely believed, so the impact was relatively small. However, 2 caveats to the Stanford work:
In other words this may underplay the total impact of Fake News, but even so the model is still showing it has to be a LOT more effective to actually swing the votes. Our view is its a marginal contributor, but in a 50/50 split election (which in effect this was) even small margins can be effective, especially if used in conjunction with a number of other small nudges.
Also, the Stanford model's definition of "Fake News" is very strict - we believe there is far more "False" news - news that bends the truth, or is economical with it - in circulation, and that acts in a similar manner. A lot of this sort of news is meat and drink to more "respected" media as well (and it is they that are leading the complaints against "Fake" news).
At any rate, expect more use of Fake News in future campaigning, and in attempts to persuade in general.
We have looked at how our systems can counter this, and believe we have some solutions
Tuesday, January 10. 2017
Tronc won the "silly renaming" rights for 2016, and we aren't long in 2017 for the first contender - Bits of Yahoo! not sold to Verizon will be named Altaba (without an ! even!) - TechCrunch:
Despite hiccups*, Yahoo’s planned sale to Verizon appears to be moving forward — but some portions of the company will be left behind and renamed Altaba Inc.
There is no truth in the rumour that Verizon will be re-named Verizon!
*That'll be the many millions of accounts found hacked in 2013/14, and the Peanut Butter problem
Friday, January 6. 2017
Deja views...the bitcoin valuation from c 2013. Graph courtesy coindesk.com
Yes, we've been tracking the ups and downs of Bitcoin value since as long as they have been traded, and after a rapid boom in the last few weeks the expected bust has set in - so far only about 12% down on it's high, which was about where it got before the big 2013 bustup
Monday, January 2. 2017
Amazon drone patent drawing
Amazon has had a patent granted to fly blimp warehouses above cities so drones can deliver your goodies from the blimps. Leaving aside the "how the hell do you get a patent for something so damn obvious and that has already been done", the question is why blimps?
The answer is the appalling logistic costs of transporting products by drone. As we showed in the previous note on this area (over here), drone delivery has some major problems:
(i) Small payload - many trips are required to satisfy even a moderately large order, a weekly shop would take 20+ drones to deliver. Heavy lifting drones are unlikely to be a feature of urban environments anytime soon, they are very dangerous if anything goes wrong.
So there are essentially two solutions to the problem - either put physical warehouses in very high cost urban land (ie buy the Royal Mail or Big Yellow Storage) which hugely increases costs of warehouse square footage, or have mobile warehouses that move in closer so reducing the back and forth distance each drone flies.
Why blimps? Simply put, our analysis shows that the sheer number of drones needed to replace vans would be tantamount to huge swarms of these devices, the noise, flying traffic and risk of having them at street level would intolerable. Thus the only other available option is to fly them up and down to hovering blimps. Downside is the blimps are aircraft and thus have to fly at a safe height (tens of thousand feet), so drones (which, remember, for this application are basically helicopters so cannot benefit from any lift generated by wings) will have to expend a lot of energy climbing
We say only available option - the other option is of course vans, travelling on roads. These are high payload carrying, can optimise multiple drops and are non intrusive as they stick to existing transport networks. Best of all, they don't fall out of the sky.
Also, as we showed in the previous note, vans have a useful device - called a driver - that can negotiate that most tricky of problems, the last yards delivery to the customer....
Saturday, December 31. 2016
It's time for Broadstuff's annual Tech predictions for 2017, or rather where we Puncture The Hype. Given our whole operating mantra is to give realistic advice to clients on new technology opportunities, getting behind the hype is essential and if we can do it a bit tongue in cheek, well that adds to the entertainment (You can see our stellar 2016 record over here). Of course it is in Listicle format as countless analytics show more people read listicles than ordinary articles, so consider yourself "nudged".
Anyway, here goes:
1. The Reality of Legacy - the weightless rise of cloud, social, analytics etc platform companies will increasingly find the gravitational pull of legacy platforms restrains them. That is where much of the data and a lot of the core processes flow, and for these new systems to move from the periphery to the centre of the enterprise they will need to interoperate with the old gradgrinds - which will force reality into all the overblown economic "savings" projected from all the new shiny products hitting your screens.
2. Cloud comes to Earth - Not only the above, but the limiting economics of the rental cloud model, the lower service responsiveness for sophisticated users, and its security risks will weigh more heavily on heavy users of computing power. Expect far more to be made of hybrid cloud + user managed services. Last year we believed Cloud had already hit these buffers and (as has been the trend) would rename itself again - - we were wrong, but will predict that 2017 is the year of transition and we'll see more Cloud - X and Y-Cloud services
3. Social, meet Regulation - the lesson of 2016 is that Social as a vector of Fake News, heavy trolling and abuse, use to communicate "ist" dogmas, a vessel of Millenial mental illness et al has sparked increasing ire (whether you believe its valid or not) in a lot of groups and organisations which do have political influence, so expect increasing pressure for social platforms to be regulated or constrained in a raft of ways. In other news, all Social platforms will busily copy each others' features to differentiate themselves until they all look the same.
4. 2017 will be hyped as The Year of Mobile, will disappoint, and 2018 will later become the New Year of Mobile - In short, we doubt the planet will be eaten by Mobile in 2017. Its days of exponential hypergrowth are ending, and it's settling down into a mature system with comfortable growth and new services largely cannibalising old ones, like the PC industry once it came to the top of its S curve. Worry about Ad-blocking will continue to grow faster than Ad-blocking, but levels of Ad-blocking will start to impact service models, driving increasing concern in the Ad industry but expect no reform yet.
5. AI / Machine Learning - hype will continue to grow to stratospheric levels while the underlying services massively undershoot the rah-rah. Actual deployment will continue to be in tight, narrowly defined verticals with manageable solution spaces, while theoretical deployment will rocket to the stars, to infinity and beyond. The major problem with AI right now though is much of what is being called "AI" in the hypewave are in fact just closed loop system dynamic algorithms, which is nothing more than a variant of...
6. Big Data and Analytics - people are starting to realise that "big + data" is not the same as "useful information".and that pure analytics is useless without knowing what to do with it. This will be the real focus in 2017 and for the next new years. AI or not, suspicion and distrust of these systems will grow and there will be a few high profile system failures in 2017 that will increase pressure to regulate or force transparency on what these algorithms are doing. Also, on the security/privacy front there will be no letup in data heists - much big data is concentrated by entities whose main concern is its exploitation, not safekeeping - so it's only going to get worse.
7. Internet of Things - the current consumer IoT wave will increasingly look like a busted flush as ongoing concerns about security, privacy and long term service viability grow owing to more data heists, failed services and pulled products. It will sink into the Gartner hype cycle's "slough of despond" in 2017. Lack of universal standards and producer preference for walled gardens means that interworking and open data transportation will remain a pipe dream. Thus Industrial IoT and tightly specific consumer verticals with simpler offerings will be where the real growth is. Wearables, a consumer vertical of IoT, underperformed hugely vs forecasts in 2016 and will continue this slow trend in 2017 owing mainly to the standards and service risk issues.
8. Robots/Mechatronics/Drones - are essentially "vertical" applications of AI/Analytics maths that create real and tangible value (the real cost of physical production essentially drives out all the spurious "Unicorn" business models that drives headline hypergrowth however). But drone/robotic delivery is being totally oversold, once the sheer volumes of these things required for package delivery and thus moving along sidewalks/flying in suburban areas becomes clear, the only hockey stick curve will be urban resident resistance. Those most hyped of robots, automatic cars, are a pipe dream for several years still. There will be high-hype trials of course, but outside of extremely well manicured environments these devices will still require human control* - so the first major deployments will be industrial (after all, that's where all the previous generations went first) and very structured and dedicated usage paths ie "road-trains/trams".
9. Remote Production (3D Printing et al) - will continue to grow, but also will not "eat the world" in 2017. Expect the real growth to be in remote production by more conventional devices like textile weaving & printing, CNC machining etc rather than deposition (aka "printing") techniques. It is still largely high cost and low quality compared to older technologies, so will remain in niches where its unique properties create extraordinary value. There will continue to be "Much Wow" headlines but big picture will be "so Small". Makers are not going to replace Manufacturers any time soon.
10. Unicorns / "New Ways of Working" / Gig Economy - 2017 will be Shit or Bust year for many Unicorns that rely on "New Ways of Working" using worker exploitation or (and, in some cases) regulatory arbitrage based business models. Many Unicorns are shitting themselves, some big names will go bust in 2017 - especially at risk are those that are growing by using investor money to buy market share that then run into regulatory headwinds. There is much flummery going on as Unicorns caught on the horn of this dilemma attempt to pivot. You can't make book on this yet, unfortunately...
And, as its now the rage in "curating" listicles, here's a Bonus Prediction:
11. AR/VR - Useful bits of AR will become integrated into Mobile and Wearable devices over time, VR will be a niche pursuit until (if) price points come down hugely and even then its not likely to expand much farther than the gaming aficionado market. Resist all blandishments that this is the future, it won't be.
*As an aside, there have been a number of deaths from hybrid auto-cars beinh used outside their "safe zones", so far the makers haven't been touched. So far....
Friday, December 30. 2016
As no-one predicted ever (erm, 2012 anyone?), hackers have hacked fingerprint recognition technologies on mobiles - ZD Net
New research, set to be announced at the Black Hat conference in Las Vegas on Wednesday, by FireEye researchers Tao Wei and Yulong Zhang outlined new ways to attack Android devices to extract user fingerprints.
The problem isn't just limited to mobile devices. The researchers warned that many of the attacks they note in their talk also apply to high-end laptops with fingerprint sensors. Vendors are rushing out patches, but its clear that this extra-security approach is hackable
And the problem is, once hacked, you can't change your fingerprint.....
Thursday, December 29. 2016
Drone Drop Economics - Copyright Broadsight Ltd
I was reading an MiT Technology article on Ocado's super-automated warehouse (h/t Renaissance Chambara) and after all the description of the robot warehouse it talked about logistics and the use of robot vans for delivery. (For non UK people, Ocado picks the food at its warehouse and then delivers to your door via a fleet of vans).
Now we use Ocado and similar, as do many suburban Londoners, as the time to drive and shop in London's traffic makes it very convenient and at c £3 per delivery, it actually almost costs in vs taking the car - or at least I and many others are happy to pay the extra cost for the saved time and hassle.
But when I started to think about Robot vans negotiating London's narrow and traffic choked suburban streets, and then doing a robo-delivery to the countless different front of house arrangements - gate / drive / stairs / entry / stairs (a lot of people use Ocado et al so the driver can take their heavy shopping up several flights of stairs to their apartment), I came to the conclusion that this was all going to be a long, long way off and the article was just doing the usual techno-hype crapdance at that point. I also then understood exactly why Amazon was interested in drones - heck, if they can fly contraband to prisoners they can deliver to Nth floor apartments (maybe in the future all apartments will have dronepads bolted on)
But then I started to think about payloads and flight costs and quickly realised that drones are probably only ever going to deliver light, high value items, and grocery shopping probably isn't in that category, because it's heavy and relatively low value.
Now no consultant can go for long without a trusty 2 x 2 matrix, so sure enough, there is one at the top of the page to illustrate the point - in essence the only drone delivery space is light and valuable (top left) as:
(ii) Payload limits - to lift heavy stuff you need big drones, but they themselves are heavy and thus will be too dangerous to have whizzing around suburban skies Right now there is a 20kg weight limit (drone plus payload) outside of safe spaces but that's because there are so few drones. A weight of just a few kilograms dropping out the skies can seriously hurt someone, so this weight limit will come down over time in populated areas, and that limits payload so your typical Ocado shop will need a LOT of smaller drones each carrying their kilo or so of goods. (Think of your weekly shop, break it into c 1 kg / 2 pound lots, and then think of the number of drones needed)
Here comes the weekly shop....
Somehow, I don't think the optimal solution is flying the 40 or so drones small enough to be safe in a big flock from out of town warehouse some 20+ miles away on a fly/drop/return flightplan.
Next step will probably be someone suggesting giant dirigibles flying above the suburbs as mobile van-hives with drones buzzing to and fro from them, but all this does is reduce drone to/from flight and replace it with a dirigible flight cost. The weight/value equation will still cut in, big time.
(Update - it appears Amazon have patented the blimp-drone thing - that was quick, I only wrote about it this morning....)
Saturday, December 17. 2016
At the beginning of the year we made ten major predictions for key technologies this year. They are tongue in cheek, as an antidote to the portentious projections that come out at this time every year but there is always a serious point behind them. Here's how we did :
1. There will be an infinite number of Tech forecasts coming out in the next few days/weeks. All will be mainly wrong (except ours of course).
Well this was an easy one to get right of course Our results were 6 spot on on, one happened this year when we thought it wouldn't, one didn't happen which we thought would, and one we got half right - so 6 1/2 out of 9, that's a higher hit rate than most so we will award ourselves a point for this one as well .
2. By mid 2016 there will be an Uber for anything and everything being pitched by someone, somewhere. By late 2016 an Ubermensch will kick-start an Uber for Ubers, and the cycle will implode up its own asymptote.
Not quite an Uber for Uber, we weren't serious about that - but there have been some real doozies pitched - my favourite is the Uber for your washing machine.
3. There will be an unseemly rush of Unicorns for the IPO gate. It will be seen that many Unicorns are just turkeys though, and they will jump the shark rather than the gate, which will mostly remain bolted.
We got this one half wrong - not in our view that that many had to rush for the gate or go to the wall - but that they could. Turns out that the bubblemarket started to deflate and there were very, very few - but there were quite a few failures and a general realisation that Unicorns were not what they should be - even the Economist at year end even pointing out that " a great many of the business models that have led to astonishing valuations for startups – like Uber – may be based on attempts to end run regulations to exploit workers, as opposed to technological breakthroughs"
4. By the end of 2016, the predictions that "2016 will be the Year of Mobile X" will be reset to 2017.
This is our annual snark at Planet Mobile, which - since we started the blog 10 years ago - has always been extremely (over)hyped about its prospects each new year. Our favourite is Ben Evan's series that started in 2014 that "mobile is eating the world". This year was no exception, but the world still remains uneaten, and next year's prediction of "mobile eating the world" has just come out. By the way, they are very good reads, but the title is endemic of Planet Mobile - no ways will you ever see a report saying "Mobile growth 2017 - doing nicely enough thank you" .
5. The worry about Ad Blockers being used will continue to increase exponentially faster than the use of Ad Blockers grows
The panic was overstated as predicted but one hopeful sign is the increasing realisation in the Ad industry that (i) they may be a cause of the problem and (ii) people who are using Ad blockers do not want to talk to you.
6. Another Tech Consumer Wearable will be released amongst great fanfare, and sink in a sea of indifference.
Well, Pebble surpassed us here and overall the whole market has still not lived up to the hype in 2016, and by some measures is even in decline. This year the Apple wireless ear thingies were released, and...well, can you remember what they were called?
7. Someone will point out that The Cloud is just Bureau Computing 2.0 and a natural commodity*, and it will be rebadged (again)
We were wrong - Cloud is still Cloud, the predicted total new-wine-into-old phase change has been pushed back
8. The Internet of Things Fridge use case will still be the most oft repeated one in public. In private, surveillance and monitoring will be.
You can't make it up - these two use cases crashed together and hit the headlines in a big way when the Mirai virus roped in surveillance cameras as well as all those dumb fridges, toasters etc using endemic IoT insecurty to launch one of the biggest DDoS attacks of the year. I guess technically we were wrong, in that the hidden use case became very public, but we are delighted to be wrong here, it as really put the cat among the IoT pigeons!
9. Artifical Intelligence will be seen as The Answer, despite the salutary experience of Die Antwoord (The Answer) with robot artificial intelligence. Someone will hack a traffic routefinder system and send all cars to Coventry.
The hype over AI has indeed gone into overdrive, but our point about self drive cars not really ready for use had more tragic outcomes than our tongue in cheek comment, with a number of drivers being killed in crashes
10. There will be multiple leaks of private user data, and users will still blink lemming-like every time it happens, believing that their free service provider values them as a special, unique snowflake (well it does of course, that value is their sale price)
And there were! Just before writing this Yahoo put the bow on the show admitting 1 billion accounts had been hacked in 2014. - and yet everyone still just seems to sign up, hit the "I agree" and dump their lives onto the 'Web.
So by our reckoning two wrong (one almost -over-wrong), one half wrong, six right, one we gifted ourself so 75% hit rate. Arguably the mobile one and "we'll be more accurate" are self righting, so 5 1/2 right out of 8 may be a better marker - a mere 69% then.....
Friday, December 2. 2016
The increasingly standard Election Prediction Method of predicting one winner till the total volte-farce at the last minute
After their superb showing in Brexit and the US Election, we bring you the best yet - the Richmond by-election in London. One candidate is definitely going to win, until the last minute when it seems the other is. (Betfair win probability graph, hat tip to Mike Smithson)
Those who don't read this blog may not know that we correctly predicted Trump's election (and had the balls to go public about it 3 days before the election), but what we never understood was what the pollsters were up to - our monitoring systems told us that Trump was way ahead from the get go. Yes there was a spike for Hillary at the end, but it was from behind.
Yet the pollsters consistently showed her (and Remain, and now Zac Goldsmith) way in front till the Shock (hah!) Ending - how come? There are really only 3 possible reasons:
i) Their methods are crap, they are talking to their choir
We originally thought it was (ii), but now we begin to wonder....
(Page 1 of 279, totaling 2786 entries) » next page
More Broad Stuff
Poll of the Week
Will Augmented reality just be a flash in the pan?
Creative Commons Licence
Original content in this work is licensed under a Creative Commons License