This thought-provoking quote by William Gibson has been on my mind recently. The frantic pace of AI development contrasts sharply with the casual indifference of friends and family who do not care about cutting-edge technology.
Most people outside the tech community may have heard about ChatGPT, LLMs, or other “autonomous” technology in passing.
However, we will increasingly see these worlds intersect. Take, for example, this amusing video of a San Francisco police officer attempting to reason with a wayward Waymo car.
The cop steps before the slow-moving vehicle, commanding it to stop and stay like an errant puppy. He then lights a flare in front of the car, hoping the smoke would make it stop.
The video is funny but is also a cautionary tale of the types of issues that we will face when introducing autonomous agents to the broader public.
Just like the bewildered cop, we will have to deal with users who do not understand the capabilities and limitations of new technology.
Designing effective User Interfaces and Experiences for these complex new technologies will be critical to broad and safe adoption.
A demo cost Google’s shareholders $100bn dollars last week. Why?
Google’s Share Price after the Bard event
Google has dominated search and online advertising for the last twenty years. And yet, it seems badly shaken by Microsoft’s moves to include a ChatGPT-like model in Bing search results.
Why is this a threat to Google?
1️⃣ Advertising: Google’s revenues are driven by the advertisements it displays next to search results. The integration of language models allows users to get answers – removing the need to navigate to websites or view ads for a significant subset of queries.
2️⃣ Capital Expenditure: Search queries on Google cost around $0.01 (see link in the comments for some analysis). Integrating an LLM like ChatGPT *could* cost an additional 4/10th of a cent per query since the costs of training and inference are high. Even with optimization, integrating LLMs into Google search will increase costs in running search queries. According to some estimates, this directly impacts the bottom line to almost $40bn.
3️⃣ Microsoft’s Position: Bing (and, more broadly, search) represents a small portion of Microsoft’s total revenues. Microsoft can afford to make search expensive and disrupt Google’s near-monopoly. Indeed Satya Nadella, in his interviews last week, said as much (see comments).
4️⃣ Google’s Cautious AI Strategy: Google remains a pioneer in AI research. After all, the “T” in GPT stands for Transformer – a type of ML model created at Google! Google’s strategy has to sprinkle AI in products such as Assistant, Gmail, Google Docs, etc. While they probably have sophisticated LLMs (see LaMDA, for example) on hand, Google seems to have held off releasing an AI-first product to avoid disrupting their search monopoly.
5️⃣ Curse of the demo: Google’s AI presentation seemed rushed and a clear reaction to Microsoft’s moves. LLMs are known to generate inaccurate results, but they didn’t catch a seemingly obvious error made by their BARD LLM in a recorded video. This further reinforced the market sentiment that Google seems to have lost its way.
In the last month, we have had huge layoffs across technology, yet the “real economy” seems robust. What is going on?
Meta is making 2023 ‘a year of efficiency’. Microsoft, Alphabet, and many other companies have stated economic headwinds as the reason for letting thousands of people go.
However, last week, the US posted the lowest unemployment numbers in 50 years(!) while adding half a million jobs.
He points to 4 factors that are causing this disconnect:
1️⃣ 😷 The COVID Hangover -> Companies assumed COVID meant a permanent acceleration of eCommerce spending. Customer behavior has reverted (to a certain extent) to pre-pandemic patterns
2️⃣ 💻 The Hardware Cycle -> Hardware spending is cyclical. After bringing forward spending due to the pandemic, customers are unlikely to buy new hardware for a while.
3️⃣ 📈 Rising interest rates -> The era of free money is over. Investing in loss-making technology companies in anticipation of a future payout is no longer attractive.
4️⃣ 🛑 Apple’s Application Tracking Transparency (ATT) -> ATT has made it difficult to track the effectiveness of advertising spending. This caused enormous problems for companies like Meta, Snap, etc. that rely on advertising.
AI is having a moment. The emergence of Generative AI models showcased by ChatGPT, DALL-E, and others has caused much excitement and angst.
Will the children on ChatGPT take our jobs?
Will code generation tools like Github Copilot built on top of Large Language Models make software engineers as redundant as Telegraph Operators?
As we navigate this brave new world of AI, prompt engineering, and breathless hype, it is worth looking at these AI models’ capabilities and how they function.
Models like the ones ChatGPT uses are trained on massive amounts of data to act as prediction machines.
I.e., they can predict that “Apple” is more likely than “Astronaut” to occur in a sentence starting with: “I ate an.. “.
The only thing these models know is what is in their training data.
For example, GitHub Copilot will generate better Python or Java code than Haskell.
Why? Because there is way less open-source code available in Haskell than in Python.
If you ask ChatGPT to create the plot of a science fiction film involving AI, it defaults to the most predictable template.
“Rogue AI is bent on world domination until a group of plucky misfit scientists and tough soldiers stops it.”
Not quite HAL9000 or Marvin the Paranoid Android.
Why? Because this is the most common science fiction film plot.
Cats and Hats
Generative AI may generate infinite variations of a cat wearing a hat, but it has yet to be Dr. Suess.
AI is not going to make knowledge work obsolete. But, the focus will shift from Knowledge to Creativity and Problem-Solving.
Just because you can do it doesn’t make it a great business model. Take music streaming, for example.
Image by Chloe Ridgeway on Unsplash
Spotify, the world’s most popular streaming service, has been the target of some Internet ire in the last week or so. Neil Young, the creator of the legendary Pono digital media player (apparently he made some music too?), decided he didn’t want anything to do with Spotify.
Why all the righteous indignation?
Spotify pays Joe Rogan, a media personality / MMA commentator / master of “doing his own research,” over $100m to have exclusive rights to his wildly popular podcast.
Apparently, Mr. Rogan has some interesting ideas around COVID, vaccinations, and horse de-worming medication. Not particularly controversial topics 😬.
Why is this a big deal for Spotify?
Music streaming is a terrible business. Spotify has been bleeding cash for years and only recently turned a meager profit. The company had an operating margin of 1.4% in the first nine months of last year. No hockey sticks in sight.
The reason? It has to pay royalties to music labels for each music stream. The value from streaming accrues to the music companies, not to the streamers or artists.
Spotify makes its money not from streaming but from selling subscriptions and advertising.
This is where podcasts come in. Spotify pays millions to Joe Rogan because he brings in a massive audience in the highly desirable 18-34 demographic. Spotify offers targeted advertising on podcasts to its most important customers, advertisers. This makes much more economic sense than making tiny margins on each stream of, let’s say, “Rockin’ in the Free World.”
The risk to Spotify in this, slightly ridiculous, situation is not losing access to rock & roll; its not being able to monetize their investments in podcasting.
Spotify would rather you come for the music and stay for Elon Musk smoking some fine herb with his buddy Joe Rogan.
They have set up expectations for their users that they can stream any song at any time. So they have to double down on more economically viable content like the Joe Rogan Experience.
I am sure there is a Neil Young song about rocks and hard places..
Roblox is one of the world’s biggest game platforms. With over fifty million daily users, it is a wildly popular platform to build and play games.
In October last year, they had an outage where the entire platform was down for over 72 hours. This was all over the news at the time..
Today, Roblox published a post mortem about the incident. It is fascinating reading for anyone interested in distributed systems, DevOps, and Engineering (link below). I will write up a more detailed note in a couple of days.
Summary – The outage was due to an issue in their service discovery infrastructure which is implemented in Consul – Roblox is deployed on-premise(!!) on 18,000 servers which run 170,000 service instances – These services rely on Consul (from HashiCorp) for service discovery and configuration – An upgrade to Consul and the resulting switch to the way services interact with Consul lead to a cascading set of failures resulting in the outage
Some Initial Thoughts – Distributed systems are hard, and the use of service-oriented architectures come with costs of coordination and service discovery – Microservice architectures do not reduce complexity, just move it up a layer of abstraction – The complexity of the modern software stack comes not just from your code, but also from your dependencies. – Leader election is one of the hardest problems in Computer Science 🙂
As engineers and designers, we need to focus on building products that have empathy and forgiveness for their users.
Software is eating the world, but as it optimizes for engagement and retention, it leaves behind confused and exhausted users.
Companies raise millions of dollars at billion-dollar valuations. With those valuations comes a drive to add new features. With the move to SaaS for everything, user interfaces and modes of interaction seem to change overnight.
Perhaps we could take inspiration from the consumer packaged goods industry.
As a new father, I have changed diapers in various circumstances. In the dark, in the park, trying to mitigate a full-on meltdown and sometimes just trying to stem an avalanche of 💩.
And yet, the diaper works as intended. Forgiveness is built into the design. I can operate it one-handed if I have to, and it gives some protection even when not used correctly. I can be confident that the design won’t change dramatically in the next iteration.
So, dear UX designer, next time you fire up Figma, think of the humble diaper, and a poor sleep-deprived dad dealing with a poop 🌋 at 3am.
Think of the mistakes a user may make and design your application to forgive them and not punish them when they make those mistakes when addled, distracted, or simply exhausted.
We tell stories and learn by analogy. A good story maps the abstract to the concrete. For the story to function, it has to fall back to a base of shared understanding.
When learning a new technology, I try to map what I am learning to my mental model of the world. I tell myself how this new technology fits into the stories I know and try to imagine what other stories I would tell once I learn it.
I have been struggling with learning about cryptocurrencies and Web 3.0.
As someone looking to explore the nascent Web 3.0 developer landscape, I keep getting lost in layers, tokens, protocols, DApps, and DAOs. Applications (technically Protocols) like Aave have access to billions of dollars in liquidity but trying to understand how things work leads to a maze of smart contracts, oracles, and tokens interspersed with more familiar words like liquidity, interest, collateral, virtual machines, etc. Like trying to make sense of a world through a fogged-up window.
Besides me being a little slow, I think the reason is that the stories are terrible. There are plenty of grand visions of censorship-resistant platforms, the possibilities of generating life-changing wealth, but these are built on self-referential and confusing foundations. Turtles all the way down.
It is early days and the world of cryptocurrencies is still a frontier. This frontier is being explored by a rag-tag bunch of clever programmers, mathematicians, financial wizards, resourceful scammers, and brazen hustlers.
But in order for a frontier to be settled, you need not just explorers but also settlers willing to uproot their lives to claim their 160 acres. You need developers to build mundane applications that solve mundane, but important problems. You need salespeople who can articulate the value proposition of building on this new frontier. You need good stories.
To stretch this tenuous analogy: we are now in the gold rush, but along with the gold rush you also need good weather and fertile land in order for the Wild West to turn into Sunny California.
Those passionate about the emerging world of cryptocurrencies and decentralized applications need to do a better job in bringing the rest of us plebs along. Otherwise, the gold rush will be over soon and all that will be left is a barren wilderness of abandoned protocols, orphaned DAOs, and blind oracles.
I am hopeful though. The infrastructure of Web 3.0 is still under construction. I hope strong, secure, and performant platforms emerge from the current Cambrian explosion of web technology. I am inspired by the likes of Chris Dixon, Balaji Srinivasan, and others who are bringing Web 3.0 concepts to wider audiences. But will it translate into wider developer adoption and mindshare? Time will tell.
I bought my first NFT today – I am the proud owner of rushiluhar.eth. It was an interesting experience. Quite similar, in some ways, to creating my first website almost twenty years ago.
Some observations:
1. You need to know what you are doing 🤔. The world of Web 3.0 is confusing and the UX .. leaves a lot to be desired. I used the most popular wallet – Metamask and bought rushiluhar.eth from ens.domains. None of these applications are for the faint hearted.
2. You need to be patient ⏳. I wanted to use the domain to point to my newly created Ethereum wallet address. This involved two separate steps. First buying the NFT (yes, each eth domain address is a NFT), and then another step to link the domain to my ETH wallet. Each step involves a transaction, and each transaction takes at least a minute to complete. And given the costs involved (see below), the lack of feedback or clarity is .. perplexing.
3. You need to be rich 🤑. ENS is run as a non profit, but you have to pay transaction fees. Which are crazy high, and change all the time. Buying a domain (like rushiluhar.eth) costs 0.001 ETH + the number of years you want to register the domain. So for 5 years, it costs 0.006 ETH – roughly $25 today. Not bad! *But* – the gas fees were (quoted at most) 0.043 ETH – almost $200! Making it my primary ENS domain involved another ~$85 in transaction fees. Ouch.
4. You need to be an exhibitionist 😬. The blockchain is public! Every single transaction is visible. If you want to laugh at me paying exorbitant gas fees for buying a worthless vanity NFT, just hop on over to Etherscan and search for rushiluhar.eth.
5. You need to be slightly delusional 🤪. The poster child of Web 3.0 (Ethereum) claims to become the World Computer. *But* current transaction fees put doing anything interesting on it out of reach of 99.9% of the world’s population. Something does not compute..
Also feel free to send a couple of ETH my way, you know the address..
Some Notes
Domain – You can buy your very own eth domain at ENS Domains but you will need a wallet, some crypto and a basic understanding of how ENS works and why you should bother.
Wallet – Metamask has integrations with a bunch of different Web3.0 sites like OpenSea, Foundation, ens.domains, etc. Its the easiest way to setup a wallet, just don’t forget your password.
Crypto – I would suggest using a reputed exchange like Coinbase to buy or sell crypto and then transfer a small amount to your Metamask wallet.
Why bother – A brief explainer of what else you can do with an ENS domain.
If you are interested in why I decided to spend a good bit of money on a ENS domain (apart from bragging rights on LinkedIn..) , check out this thread by Balaji Srinivasan.
Benedict Evans, as usual, provides an insightful view of the “Digital Transformation” story. While crypto, machine learning, NFTs, and drones may generate the most headlines, we are also in the midst of a generational shift in how we do business. This shift is happening in boring corners of the B2B Enterprise software market but will have an impact bigger than some of the other, more alluring, technology trends.
Companies like UiPath (process automation) have successfully targeted the dull areas of enterprise software that are ripe for automation and streamlining. The software in “software eats the world” may include headline-grabbing items such as machine learning and distributed ledgers. But, much more significant changes are being brought about by the adoption of SaaS applications and workflows. Twenty years ago, you couldn’t add a new software application without going through procurement and IT. Now, all you need is a corporate credit card.
Demographic changes and shocks such as Covid are only accelerating the technology megatrends for which “Digital Transformation” is a catch-all term.
The concept of “generational shift” also works on multiple levels. Prolonged and painful migration projects can last for a generation (or longer). But we also have an entire generation of programmers and systems administrators who are now retiring, and you can’t find the talent to keep workhorse systems going.
I am thinking about the second-order consequences of this shift to a software-first world. There is going to be more efficiency, more competition, and a chance for aggressive upstarts to ride the technology wave and displace (rather than disrupt) less agile incumbents. But, we will also have a generational loss of knowledge that cannot really be replaced by software.
As every aspect of our economy is driven by software, we start seeing some of the characteristics of software show up. We have seen shortages of food, supply chain issues across industries as well as an incredible increase in ransomware attacks as overly optimized systems break under unexpected disruptions like the Covid pandemic, or insecure systems are targeted by malicious actors.
So, while I think this trend towards more Digital Transformation is good – in aggregate; there are also serious consequences that we may not being much attention to as we continue to be driven by software to optimize.
