Whilst in Arrakis

The Future Will Be Better Tomorrow

Sunset on Intel


the transition away from the x86 architecture is a discussion that I have heard all my life. Even back when I started it was considered sub-optimal and needed a rewrite from the ground up, because the demand for DOS/Windows was skyrocketing, it could not be attempted. And this problem of a shaky foundation remained for literally decades. While at the same time Intel lagging behind in terms of die shrink, and performance gains not due to architectural advancements but more processors, more speed, and stuff like speculative execution (which ends up causing unfixable security issues). Basically there was a performance wall that x86 was going against, and they were brute forcing it.

The market dominance was their biggest advantage, but also the biggest barrier to developing a leaner alternative.

In new markets, they did develop an alternative. As server computing grew they developed Itanium, which was IA-64 and not x86, but its development was based on old assumptions (or tests, depending on who you ask) – meaning that it was optimized for the previous generation of supercomputers, which were mostly about scientific computation. But the server market that was growing was cloud computing, where IA-64’s performance just was not there, so they ended up releasing an expensive server processor that would not perform better than their previous architecture, and that anyway the x86 didn’t want. Itanium was finally killed last year.

Around the mid-2000s, for reason known only to them, as the market for smartphones was booming like crazy, they decided to sell their mobile chipset division (again, “mobile” meant “how to squeeze PC performance and power consumption in a phone” rather than developing a new product designed around the constraint of a new device).

Better/leaner/lower power consumption alternatives were actually being funded (Transmeta being one), but because they targeted an existing market, they all faced the headwinds of developers not wanting to recompile programs for something that would ship 100,000 devices at best. With the release of the iPhone this changed, their ARM architecture got better at every cycle, until reaching the point (and this is MIND BLOWING) of outperforming x86 in emulation mode.

I suspect (this is a huge speculation) Jim Keller left when he told Intel that the only solution was to ditch x86 and write something else.

The monopoly days of Intel CPUs in desktops are over. Microsoft adjusted from going to ~90% computing market share to ~8% by moving onto the cloud, and they have done it extremely successfully. I don’t think Intel can pull the same feat.

The biggest issues of Intel, in my opinion, were:

No funding external innovation.

they did not invest in the competition. Finding a VC backing a silicon startup was very hard in the Bay Area. Ironically it was one of the areas where more funding was available in Cambridge (which is almost never the case)

Not investing in mobile

decided against investing in R&D to win the iPhone business, because it was too expensive. Around 2005-2010 (IIRC) there were about 750m PCs, and 2bn phones. PC shipments were slowing down for the first time in history, and it was clear that all phones would have had a dual processor in a decade. The era of PC was not over, just that mobile was going to be a bigger market. By then it was clear that Itanium was not going to be a solution for commercial cloud computing, web hosting etc.

Getting out of the GPU gaming business

This was harder to plan, but (to some extent correctly) they got out of gaming. Alas it turned out that a lot of problems use linear algebra besides gaming (machine learning, crypto), and it was clear by early 2000s that having some sort of programmable GPUs could offload computation and speed up programs massively. In the early 2000s programming for it was clunky, so can understand the skepticism, and even in 2010 it was not obvious how big deep learning was going to be (ironically not even to NVIDIA themselves)

TLDR; none of those can be reversed in time now that their monopoly is over. A company like Intel is huge, they are a key national defense supplier, so they can keep going decades, but the decline is unstoppable.

A Personal Star Wars story

I never really pestered my kids with Star Wars, just because it was a huge pillar in my childhood doesn’t mean that they care about it. So was pleasantly surprised this week when my youngest asked me to have a SW marathon, in preparation for Chapter IX.

I think his biggest influence in this decision was how much he enjoys watching The Mandalorian.

I have not really watched A New Hope since the 1997 cinema re-release and certainly had never seen it in 4k, and wow, even something that I know so intimately gave such powerful goosebumps. Rarely I get to relive elements of my childhood that stand the test of time.

I just finished the first one, and really glad I did.

Anyway here’s the planned list in order:

Star Wars Canteen (LEGO version, obviously)
Rogue One
Xmas Special (only the animated segment)

Heroes never die


Was digging through some old folders and found a couple of documents that might be of interest. These are very early documents of our work on EverQuest Next. At the time we wanted to do AI storytelling but Sony was clear that they were only interested in combat AI and that we should stick to it. These were our proposals and we did other work, with a much bigger scope after that.

The first document is the initial pitch after we were told the scope of the game that is now public (at this point we hadn’t seen anything about the game) and it wasn’t clear how many features we would have to develop for the final product. As you can see Magic (the card game) was a huge influence on the prototyping stage.



The second document was our proposal for a demo of the AI combat system, and that was what helped us advance the discussion for our involvement in EQN. It’s pretty dry but page 2 gives you an idea of how we felt about the problem, and that adding AI combat for us was just a small step in a grander vision of an AI-driven game.



Have some more documents but the earliest NDA with SOE I have found was signed after these documents were produced. I have done my due diligence and if there is a disagreement please produce the documents that show otherwise.

A lot of people contributed to this, mostly Stéphane Bura, Dave Mark, Sam Baker, and Brian Green.

I have several non-EQN related documents about Storybricks but re-reading them I get really embarrassed by the naïveté I had at the time related to game design (the AI was surprisingly spot-on) so not really keen to share. Others were game pitches of stuff that I or Stéphane might do one day (we built a lot of concepts) so no dice.

The design for the slides was shamelessly ripped off from Tim Cain.

Again want to stress how great is to use Magic to prototype game ideas. Try it.

Good old OnLive

You were the gift that kept on giving. Raised $200m at a valuation of $1.8bn:

July 2011: Cloud-based video games will put pirates out of business

July 2011: OnLive CEO says his pet wireless project has sub-millisecond latency

January 2012: OnLive can turn your iPad into a remote desktop with live-streamed MS Office apps and more

January 2012: Review: The futuristic OnLive Desktop runs Windows apps on the iPad

February 2012: OnLive delivers ridiculously fast web browsing on the iPad

March 2012: OnLive brings its amazing Windows desktop access to Android tablets

October 2012: OnLive’s assets were sold for only $4.8 million

March 2014: OnLive is reborn with IGN founder as chief and two new cloud-based gaming services

April 2015: OnLive will shut down its cloud gaming services and sell assets to Sony

(all these articles were written by one outlet, the only one that would give extensive coverage to OnLive and provide good reviews despite any absence of customer traction or any questioning of the company fundamentals)

Snapchat is now augmented reality-first



Facebook did the deal with Oculus mostly so that they would own the post-mobile hardware platform. PCs are the old platform and Wintel won, current generation platforms are dominated by Google Android and Apple iOS, and it’s unclear if there is room for a third competitor at scale[1]. Facebook decided against making their own mobile phone and instead invest in what was coming next.

So Facebook acquired Oculus.

Virtual reality was always a stepping stone to augmented reality[2], as today you need to find time to actually use a VR headset. As the chart below shows, attention is a very fought-over currency.


This is not the case with augmented reality, as you can double stack minutes doing other stuff and still consumer AR content (via a pair of glasses for example[3]).

Snapchat is rumored to be tossing their hat in the space and go for an AR play but I would argue that Snapchat is an augmented reality company. In the same way Google pivoted into an artificial intelligence company around the time of the Deepmind acquisition [4] Snapchat started with self-deleting messages but their lenses [5] have clearly taken over and their are driving the virality for new users. If you take that experience further then to me the mission of Snapchat becomes clear: it’s the lens through which you experience the world. A better, more beautiful, and augmented world.




Once you acknowledge that, there is no point in limiting it to a phone. Instead use lenses to mediate interaction with reality making a pair of glasses the perfect fit [6] and this is when it gets really interesting.

There are other companies in the AR space but they don’t have the passionate user base of Snapchat or they focus too much on games, which is cool and a big market in itself, but entertainment is tiny compared to say communications or commerce. A Snapchat reality lens powered by artificial intelligence and computer vision would be silly, incredible, and have wide adoption all at the same time.

It would also be a  great developer platform with clear ways to monetize. Expanding the lens to incorporate objects different than faces (e.g. cans of soda, so if you are looking at a Coke can it shows some cool animation) or different input lenses (e.g. an audio lens that generates custom clips when something is played or replaces ambient sound – possibly done with noise cancelling headphones). Like lenses today it doesn’t have to be dominated by advertisements blocking your reality.

Oculus is essentially a game console manufacturer, and it’s not going great with hardcore gamers. The product caters to gamers who can spend big 💸💸💸 on a new PC, but it can only be played for a bit before getting a headache. The design is not quite there yet and will realistically need several iterations more. And the locking down of games is pushing consumers and game developers away from it and onto the HTC Vive [7].

Here’s what I believe is going to happen:

Facebook’s mission is simple and it’s about sharing experiences and content with others. To date Oculus doesn’t really fit that mission and possibly never will. Because of this I expect Oculus to be spun off or sold to a game company[8] or potentially to Microsoft as HoloLens is not a credible product (or even a product) and Oculus could be part of the Xbox unit.

Snapchat however is a fit and also a strategic threat to Facebook already (it’s half the size of Instagram). More importantly its users are all about sharing. So it could fit the vacant space of AR left by Oculus. It’s well known that Facebook tried to buy them for $3bn but the offer was rejected, eerily similar as to when Google tried to buy Facebook for $15bn [9]. Hopefully Facebook will remember that.

Snapchat’s self deleting messages are no longer the main reason why people join, it’s about AR and messaging (with AR). So expect them doubling down on lenses and possibly become a full augmented reality platform with their own hardware. But here’s the big difference with other AR vendors: they are all either stacked into enterprise edge cases (Google Glass at Work) or games. Snapchat is different. Snapchat is Minecraft. It’s simple rules and (mostly young) people have appropriated and explored and created their own language and culture on top of it. It has that new corinthian leather feeling that Facebook used to have. I believe this culture is Snapchat’s biggest asset.

This culture will adopt a new “toy” in the form of AR and go with it, while other competing platform will struggle by trying to meet the expectations set by their bullish hype.

TLDR: Snapchat is an augmented reality-first company, Facebook will divest Oculus and instead try to acquire Snapchat.


[1] Even though there could be an opportunity for a Cyanogen, or a new OEM to conquer critical market share by making it easy to install an OS on a phone. Blackberry is probably toast and ready to be absorbed by some large company like HP.

[2]  But solving VR means solving a subset of technical problems and it can be easier that going AR native.

[3] Even though it could give life to pretty weird scenarios like this “Hyper-Reality” one https://vimeo.com/166807261. Although my point is that they want to make it fun for users to engage with brands and will have the brands pay for it. If the users are not engaged they won’t use the platform and therefore no critical mass for ads – therefore users > advertisers so not expecting weird stuff to happen.

[4] Yes Google was always an AI company but not really. It was a giant search engine with ads. That process was already under way but the Deepmind M&A was a clear demarcation point and right now we are seeing the results. See ‘Inside Sundar Pichai’s Plan To Put AI Everywhere‘.

[5] It’s the video filter that makes the silly faces.

[6] Glasses today and contact lenses in the future.

[7] This is the first time in more than a decade anyone is actually excited about HTC products. Incredible.

[8] I really don’t know who could afford this beyond actual console manufacturers. Plus game studios have no idea how to manage hardware. The other alternative is that they keep Oculus and starve the unit of resources until it’s small enough that can be folded into “Facebook VR Games”.

[9] IIRC can’t find the article, I think it was Megan Smith that was tasked with that directly from Larry Page.

Magic mirrors and AI

If you haven’t yet, read My Bathroom Mirror Is Smarter Than Yours do it before this post.


Smart mirrors show consumers want a multi sensory experience. Touch, vision are primary drivers of Human-Computer Interaction. Voice is a novelty that can build on top but asking people to change behavior in HCI without a clear benefit is madness.

So far only benefit of voice is while driving, people who don’t know how to write or people with disabilities. Clearly not growing markets.

Using voice we can’t receive complex data structures like lists, charts etc easily and the bandwidth is an issue.

Reading 250 words per minute
Listening 150 wpm
Speaking 105 wpm (this assumes the AI will understand 100% of the time. Sci-fi)
Typing 33 wpm

Compare the reception of all the Nest / Echo etc products with the smart mirror. That guy accidentally discovered a potential product market fit with a Medium post (world first?)

Soon we’ll see more smart mirror equipment coming for sale, DIY kits, APIs. Citymapper just released an update that would make sense on a mirror (tells you if there is going to be a delay on your route before you ask for it). In fact while the Apple Watch is a “personal” device, a lot of the apps there might make more sense on a mirror.

I think there is an opportunity to exploit multi sensory AI and nobody has cracked that, I also believe that glass as a material will have an ever increasing role in our lives (Corning mentioned using Gorilla glass for windscreens for example, and there are transparent solar cells coming to the market.)

If you are an entrepreneur looking to build magic mirrors, or a VC looking to fund one (and you should) I’d love to discuss the AI behind it.

20 years of work


Tomorrow it will be 20 years since I started my first company (21st October 1995). Hell of a ride and entirely unforeseeable.

It’s only fitting that this evening at 11pm I have received the weirdest email of my career. Literally out of this world.

I love my work. It is often a harsh mistress but the rewards are spectacular.



Sam is a smart man

These are the two of the most important tweets I have ever read in regards to business and startups:

Filter your world through this lens. Especially government work. Just because we can do something with brute force at the city/national level today does not mean it can scale at the humanity level. We need to think at that scale and suddenly only having this mechanism being viable financially can work.

But there is more. What does this mean for government policy? Do we need a National Startup Foundation that gives out seed grants? Is there a Nobel for business? I can see Cook and Zuck getting it. Do we need special education like a Msc in Lean?

It’s harder than you think (thoughts on Slack)

It’s harder than you think – Public Enemy

Slack just did a massive $100m financing round. Its parent company, Tiny Speck, started in 2009 as a game company. They did 3 rounds of funding for about $15 mil before ditching games altogether and entering the corporate IM space.

Monetizing games is a fight against entropy. They peak in two weeks and then you have to continue pushing out content, promotions, acquire users (as long as LTV > CAC and in a world where LTV goes always down and CAC goes up). If you look at the user activity curve of all successful games it look always the same. There are very few exception and these tend to be very valuable (World of Warcraft, Clash of Clans).

Successful tech companies OTOH have a curve that compounds over time. This is because providing entertainment is a fickle proposition and is due to end at some point in the near future, while solving problems or enabling solution is potentially boundless.

TL;DR raising financing for games is easier harder than tech because the adoption is very different and over time one returns more money than the other. Also you price game companies using present revenues while tech companies tend to be valued on growth and future revenues (as revenues can be built later – see ‘Business models are a commodity‘).

PS. I really <3 Slack. If Yammer was still around as a private company they would have been clobbered. Now that they are Microsoft they are going to be completely wiped out.

The problem with Siri

Dear Siri,

it’s not me, it’s you. Let me give you an example…

“Siri, I would like to book a hotel for tonight”

Here are the list of nearby hotels” [List hotels based on a Yelp search, no indication if any of the hotels has actual availability for tonight. Clicking on the list takes me to the super shitty Apple Maps. Clicking on the hotel in Maps lets me call the hotel]

(The first on the list is the Radisson) “Siri, book me the Radisson”

Here are your Bing search result for ‘raisin’

See? I have millions apps that can book the hotel already installed on my iPhone. I don’t want to call the hotel, thank you. Chances are that if I am talking to a piece of software means that I prefer doing a transaction with another piece of software and not call a hotel, figure out my corporate rate, my credit card number, my frequent flier thingie for hotels. All this data already exists in my phone and in my email account. If not, ASK.

On top of this our relationship is purely transactional. Once you spit out your result you forget about me. Me. Your user. I mean have you ever seen Tron? Users are important! And you need to remember the past few seconds of interaction to understand if the queries are linked, such as it was in this case.

85% of iOS users have never used Siri and the main reason is that it is a gimmick that does not solve any real problem. Yet.