Reviewing “18 Geeky Predictions — 2017 Edition”

Published in

Point Nine Land

7 min readDec 18, 2017

Last year it was the first time that I published my predictions and reviewed them publicly. I enjoyed the intellectual exercise of thinking over and subsequently reviewing them. So here I am again, this time reviewing 2017, just before writing about 2018.

It seems like this time I was more conservative, because I got several more right (10/20) 😜

Following the same format as last year, here we go:

#1 Security

[🤓 correct] Everybody will get hacked: governments, cities, business, apps, SMBs…
We will learn to live with it :)

That was an easy one!

What’s more interesting is that not only do the number of incidents keeps growing, but also the severity of them: in just the first half of 2017, there were more records breached than in all 2016.

Apparently, nobody is safe, the list of targets covers everything — critical infrastructure providers? ✅ Unicorns? ✅ Democracy? ✅

In some cases, there weren’t major consequences; in some others… 📉

The stock of EQUIFAX went down by almost 50% after the data breach

[🤓 correct] More investment will go into private/corporate (cyber)security. We won’t trust (anymore?) the security provided by the previous entities.

That was another easy one! 🤔

Cybersecurity spending has been growing for the last decade— in 2017 it grew up to $85bn! There are many opportunities for startups, pitch us, pls! 😇

What’s probably more interesting is to see that compared to the AWS bill:

[🤓 correct] Data governance & privacy & big brother will be a big topic, again. WhatsApp, Alexa, IoT, Machine learning… all those things are going to clash with regulators

GDPR, GDPR, GDPR …

https://trends.google.com/trends/explore?date=today%205-y&q=gdpr

#2 Blockchain

[🤓 correct] Bitcoin will grow. We will still not know why

Compared with last year, this year growth is exceptional! But with some predictions mentioning +1m USD / BCT— who knows if this is still cheap?

Sources: Pantera Capital and http://www.cryptocurrencychart.com/top/100

[💀 fail?] Many Blockchain projects will be abandoned. 💀

There are some high profile projects that are under-delivering, but I don’t know how many projects have been abandoned. To be honest, today, I don’t think this is an important question. If ICOs are going to follow a similar distribution as VC, it will be a power-law, where the only thing that matters is to be involved in the successful ones.

[💀 fail?] My bet is that the geekier ones will get traction 🚜 — ie. security certs, distributed storage, etc. — we should invest in them!💰

There are already some contenders for the main layers of technology infrastructure for the decentralised world — kudos to Alexander Lange for putting together this map.

In terms of financial success, at the moment some of those tokens are underperforming, while others are doing pretty well (1, 2, 3). But in terms of adoption it’s too early to tell. Ethereum with +30k nodes (and growing) in the network seems to be the most active one. Most of the others are still in alpha phase.

[🤓 correct?] An “important” ICO will flop 📉.

To be fair, Tezos is still pre-floating, but it seems to be messy there.

[💀 fail] But there will be a new very successful ICO (new Ethereum?) 📈

Not really. Or at least, very hard to see through the noise:

[😮 missed 😱] ICOs are becoming a challenger to VC funding

#3 Dev Tools

[💀 fail] 📢 Public Announcement: 2017 will be server-less year

😒 meh! Not so much excitement! It feels a bit “incremental” and the incumbents are already aware that this is coming, so it doesn’t seem to be creating great opportunities for startups.

https://trends.google.com/trends/explore?date=today%205-y&q=serverless,graphql,microservices

[😒 meh] Big businesses will be built by encapsulating business processes into APIs. Developers will keep building the software to eat the world, and they need APIs for that!

The last round of Stripe in 2016 was at 9bn+, but Twilio is pretty much flat since the IPO and Sendgrid is just below the 1bn mark — 22nd largest IPO of the year.

https://docs.google.com/spreadsheets/d/1wusrxoTvnag8nGOJNmulcz_M4Sa06RWT-FgQr8T5KpI/edit#gid=0

In our portfolio, Algolia and Contentful made it to the headlines by raising $53m and $28m respectively. Well done! 🤑

[🤐 dunno] Every developer will have a side SaaS business :)

It’s hard to get data on this.

One proxy could be to look at the results of Hacker News. There, the number of articles about bootstrapping or side project topics didn’t grew much (400–600 a year, 2016 and 2017). In google trends is also a pretty flat trend.

Anecdotally, as my colleague Clement Vouillon pointed, we have seen recently a new trend of Non “VC Compatible” companies.

#4 ML, AI

[🤓 correct?] ML will get commoditised.
No researchers needed. We will see it everywhere

Definitely, it’s in every pitch 🤠.

AWS et al are “democratizing” the access to ML, but some problems still require highly technical skills.

[🤐 dunno] The “war” in ML will be about training data.
“How do we acquire more for initial training?” “How do we create more?” “How do we tag it in a smarter way?”…
btw, now that people will stop thinking about improving ML algorithms, somebody might win by doing better algorithms…

I will not call it “war”, but startups with proprietary data — or smart ways to obtain it—get extra points from investors. 😘

There’s also good theoretical progress to reduce the amount of initial data required. For example, Deep Image Prior uses the initial image as the training set and AlphaZero Go uses scenario simulations to eliminate the need for initial training data.

*Image restoration problems without train/pretrain datasets*

[🤓 correct] We will invest in at least one Vertically-Focused-ML-Startup® Or an ML-First-Company™

Yes, we did it! Congrats to my colleague Louis Coppey for finding calldesk.ai

#5 Hardware, IoT, Drones, AR&VR

It will be an Alexa & Google home year. That will have two implications:

[🤓 correct] (a) It will be a Voice-Powered-Anything™ year and

Amazon already sold over 20m echo devices — as a benchmark, Apple is shipping around 200m iPhones a year. That’s not the only impressive stat, the number of skills available for the Echo — the equivalent of apps in the voice world — is also growing very fast to +15k available.

This seems to be the beginning of a great new platform. If you’re building something into this space, my colleague Savina van der Straten has been writing about why we are so bullish about voice as the next platform here, here and here.