This piece initially published April 2018

Almost everything these days has a circuit board inside. Tempo combines machine learning and robotics to crank them out faster and error-free.

Whether you’re building cell phones, drones, security cameras, rockets or driverless cars, you’re gonna’ need a circuit board to make it run. Traditionally, you would send out your design to a mom-and-pop shop to get the boards made. You’d call them up on the phone, wait 48 hours for a quote, then exchange an average of 35 emails and calls to sort out the process. After all that, you need to wait three weeks for them to hand-assemble your board and deliver it to you. After unpacking, you’d find a problem. Your engineering team would need to figure out what went wrong, then fix the design, call the circuit board maker and start the process all over again.

Tempo is changing that month-long nightmare...

Lux’s investment in Blockstack’s Series A and Token Offering

This piece initially published Dec 2017
I first heard of Bitcoin in 2013. While today’s macro political environment throws one huge story after another, in 2013 large black swan events were still … well, black swan; rare. One of these events had just occurred: Over levered banks in Cyprus banks had rapidly declared insolvency amidst a global recession. The EU and IMF were racing to structure bailouts, including freezing residents’ accounts and tapping into them as part of the process. Cyprus residents had lost faith in the local currency, Euros, spurring large purchase orders (from both domestic and globally aware international buyers) of a little known digital currency: Bitcoin. The bailouts in Cypress led to a large spike in demand for Bitcoin leading to a jump in price leading to mainstream press, for the first time...

This piece initially published in Oct 2017
Complexity rears its head across startups and Fortune 100s alike. It’s unbiased in its applications, and often crippling in its outcomes. The consequences of the future are coming at us faster than we once imagined, and in more unpredictable ways. The increasing asymmetry of opportunities and threats are hidden in exponentially growing, often silo’d, almost certainly unstructured, data. It’s almost painfully obvious: real world data’s sheer volume and dimensionality render it beyond human grok-able comprehension. The solution? Machine Learning and AI that closes the gap between ever increasing - ever complex - data and, at best, linearly scaling human horsepower. Technology that erases the chasm between data and decision making.

Bridging the gap between exponentially growing data, and linearly scalable human resources

That is precisely what...

Recursion Pharma raises $60 million to scale its Artificial Intelligence and Automation Platform

(this piece initially published Oct 3, 2017 here)

Recursion Pharma, a company we are lead Series A investors in, today is announcing an exciting and oversubscribed Series B of $60M coming from top investors in tech, biopharma, and around the world. In many ways it’s a reflection of all the company has accomplished in the last year (more on that in a bit) and fuel for growing ambitions in the rapidly coming future. Moreover, it’s a validating moment and a nod to a strong surge of entrepreneurial technologists we’ve seen and continued to fund - beyond Recursion - that have taken lessons from automation, data, systems engineering, and AI and applied them to a large industry historically seen as “outside the purview of Silicon Valley IT”: Bio Pharma. I expect Recursion’s early success to be...

Doubling Down on Clarifai

This piece was initially published on Oct 25, 2016

Machine learning, AI, Conv Nets, Deep Learning, and Neural Nets … all rapidly maturing Artificial Intelligence technologies that have simultaneously become household jargon in the Valley.

Tesla’s self driving car, Amazon Alexa, Google Search, Facebook tag recommendations, Microsoft Cortana, and Apple Siri … all novel products leveraging the above mentioned AI technologies, developed by large tech mainstays, and increasingly popular nation wide.

Technocrati cocktail banter is developed and largely kept in-house by large technology incumbents to develop new products and disrupt adjacent industries. We’re witnessing first hand the world bifurcating, the haves and the have-nots. That said, historically a rapid rise and maturation of a new technology germinates within the the confines of a select few labs...

Announcing our Series A partnership with Recursion Pharma

This piece initially published on Oct 3, 2016 at Lux Capital

A little over two years ago I first started writing and working on an ongoing thesis of how the emergence of big data sets, automation, connectivity, and machine learning is fundamentally altering how we do science — in essence disrupting the scientific method. In short, these developing technologies and emerging sciences have enabled scientists to shift from hunting and “intelligencing” analytic truths via reasoning a priori about causation to relying on reproducible and large data sets to correlate input and output — first rigorously understanding correlation and only then asking about causation. This transition not only takes the burden of the “eureka” moment away from the “lone” scientist, but shifts the intellectual responsibility increasingly so on machines...

Initially published Sept 2017
Barriers to creation are falling. Creativity is, in essence, commoditizing. This sounds hyperbolic, but a quick look at Google’s recent work with DeepDream should quell most doubt. The computer generated, yet deeply human, art generated by Google’s autonomous artist doesn’t live in a vacuum. Behind the scenes were countless libraries, SDKs, servers, containers, algorithms, and frameworks that supported the rapid iteration and creation of DeepDream, which itself was creating art de novo.

This isn’t a new trend. A few decades ago putting up a website required large capital investment, custom servers, and took months, if not quarters, to get an early iteration live. Now with AWS, mature web frameworks, and a plethora of data repositories ranging from SQL to Hadoop this same process is measured in days … if not hours. The history of software engineering has...

Context

Jake and I found ourselves watching New Girl and solving math puzzles last Friday. The. Good. Life.

Problem:

Solution

Alice and Bob each pick the index that has the first 1 in their respective arrays.

Intuition

For each permutation length, Alice and Bob nail one case where they necessarily win.

Example

Bob’s array: 001011…

Alice’s array: 011100…

Bob’s guess: 3

Alice’s guess: 2

Note: In this case they do not win, as the second number in Bob’s array (0) is not equal to the third number in Alice’s array (1).

Python script simulation

import random

this creates the randomized bit arrays
def makeBitString(bitlist, length):
    for i in range(0, length):
        bitlist.append(random.randint(0, 1))

this is the guessing algorithm. The guess output is the index of your first 1
def guess(bitlist):
    index = 0
    while (bitlist[index]!=1):
        index = index+1

Today is about [Clarifai](clarifai.com), it’s about an ambitious entrepreneur in Matt Zeiler, it’s about a true 0 to 1 step function breakthrough in technological ability, it’s about the reawakening of the New York deep tech scene, and as a least important bit, it’s about Lux’s investment alongside USV in Clarifai’s Series A.

Before going further, it makes sense to explain: How did we get here?

At Lux we’ve been investing within the intersection of deep tech and contrarian for over 10 years. The thesis and conviction to be differentiated and shoot for true outsized venture returns has led to investments in nanomaterials, electric grids, drones, digital pathology, and even satellites. Along the way we’ve questioned what it means to be VCs and rolled up our sleeves to play active hands in directly founding companies in Nuclear Waste, semiconductors, and even biotech.

However, time...