Individual “point” questions are typically used in tech interviews (“please write pseudo code to reverse a linked list on the whiteboard”), but these have a downside of coming from a fixed (though large) corpus of questions, typically which have leaked to an extent unknown to the interviewer. This allows for effective “interview prep” where prospects who brute-force-memorize reasonable answers to the corpus will get the questions “correct” and pass the interview. This will select for dedicated employees with good memory, but may unfairly punish those with reasonable breadth and depth of knowledge but who just got “unlucky” in which questions were picked.

Any test on which an applicant gets a 100% has failed to fully assess the applicant’s capability. We need a distribution where it’s basically impossible to get a perfect score to have confidence we know how to rank an applicant! So a technical interview with three memorizable answers on which the applicant gives three correct answers is discarding a lot of signal about their capabilities on the one hand and letting luck and memorization play too large a role on the other.

So the goal is instead is to do a Monte Carlo search of someone’s large knowledge space in a time bounded way, probing beyond an applicant’s “reasonable knowledge”. For a good Infinite Depth Question, the size of the Complete Answer and the random nature of the probing make “memorizing the answer” impossible. If there are areas the applicant wants to explain in more depth, they should be allowed to do so but not to wholly derail the interview, as this can allow for memorizing an answer. The goal is not to have the applicant replay a carefully memorized script!

The interviewer should have “unfairly deep” knowledge about some parts of the question that they can probe on — not to “verify the applicant has the same depth” at all, but to observe the way the applicant handles being pushed over the threshold of what they likely know or could memorize. It’s polite at the beginning of the interview to mention that some parts of the interview are explicitly going to go well beyond what is expected to be known so the applicant doesn’t panic at being presented a question to which they don’t know the answer.

One example of such a deepening probe could be around. DNS. An applicant might know that DNS sends a request to a nameserver and gets an IP address back. If we stop there, we have not probed their limit of knowledge and can’t assess how they handle this. So we proceed. How is it decided which nameserver? What kinds of name servers are there and how might they be involved in answering the request? Is the request over UDP or TCP? (An expert applicant might be expected to push back on this as a false dichotomy since other DNS resolution techniques exist, like DoT and DoH, and explain what those are and why those are interesting/useful options.) What port do DNS queries go to? What is the structuring of a DNS request packet? Response packet? What are some common DNS record types we might expect to see and what is their function? What are problems with DNS that limit its usefulness? And for forth.

Some examples of Infinite Depth Questions (there are many!):

• What happens when you turn on a computer?
• What happens when you launch a program?
• If I type in a URL into a browser and hit enter, what happens?
• What’s the difference between a VM and a container?
• How are files stored on a computer?
• If I send a SQL query to a database, how does it go about answering?
• How does WiFi work?
• How would you build a datacenter?

The Metaverse is not a bad idea not because it can’t be built. It’s totally doable, though challenging with dozens of very complex and exciting risks that require solving for: human perception, haptics, foveated rendering, extremely lightweight optics and compute, low-latency network interconnects, and edge processing of a massively sharded but contiguous world. The depth of rich and exciting components make it an absolute field day for technologists who enjoy inventing the future and who are very good at their craft.

My first real post-college startup job was at There.com, where we were building the Metaverse; one of the other employees even had “SNWCRASH” or similar as his license plate. I built an MP3 jukebox that streamed audio from a Shoutcast server and rendered the audio in 3D so you could use your ears to walk to the nearest party. It was fun. The company cratered so hard one of the other engineers on the project, Eric, founded The Lean Startup movement to make sure nobody else ever wrecked a company that hard in that way again. (Not that I think Eric’s against the Metaverse, just against companies being 5+ years in super stealth mode.)

We shouldn’t build the Metaverse because that universe is one in which people would be unhappy.

I don’t think most people want to be sitting by themselves all day long with a headset glued to their face, having virtual experiences, going on virtual tours, spending virtual money, and playing virtual games. I think real human beings made out of meat enjoy doing things that social-oriented living creatures enjoy doing: walking in nature, sharing food with friends, feeling the sun on our skin, breathing fresh air, getting good sleep, hugging and kissing and high-fiving and tackling each other. These are experiences of being present in life and that lead to happiness and fulfillment.

We’ve had about a decade of running a grand experiment of gluing a computer display into the hands of a meaningful fraction of the world’s population. There are inarguably many amazing and magical and wondrous things this has enabled, but I think it’s also hard to look at the impact that has been had on the whole — especially on childhood — and come to the conclusion that gluing that screen to our eyes would lead to a net improvement on the human condition.

You’re not going to wear a VR headset around others. It’s just awkward. You can’t make eye contact and you have no idea what the other is experiencing. So you’re not going to wear it on the couch with others in the room. Or on the bus. Or in an open-plan office. You’ll wear it when you’re in a small room, alone. This limits the utility of VR to a handful of point applications of brief usage.

AR is a natural counterpoint to this concern, but has some other issues that don’t seem to be well explored. AR is “rude” because of the information asymmetry it engenders — you and I no longer have a shared experience of space and time. If you’re looking above my shoulder, maybe you see a beautiful bird, or maybe you’re reading your email. In real life, we follow each others’ gaze as an indicator of where attention is being spent. Meaningful eye contact between speaker and listener is a strong signal of closeness and focus; it doesn’t seem likely that technology will change this fundamental instinct. So someone wearing a headset and looking around may be referencing material we’re talking about to bring up in conversation…or maybe they’re playing a video game where dragons eat my arms off. Who knows? It’s like having your phone out and staring at it with every conversation.

Social acceptance is critical for new device adoption and yet seems imperfectly accounted for. Google Glass failed not because the cameras weren’t high quality enough or the processors beefy enough. Glass failed because others in a room didn’t want a camera recorder pointed at them (even when off). It disturbed social norms and therefore people wearing them became “Glassholes”.

Conversely, fitness bands (and later Apple Watch) succeeded because having things around your wrist has been socially acceptable for a while; while glancing at your wrist may be a sign of mild distraction or boredom, I’d argue that it’s a lighter signal than pulling out your phone to go scroll.

Technologies that want to be socially accepted will have to fit into social norms of how humans want to interact with each other. Technologies that are Good will nudge people toward experiences known to make humans happy and healthy and fulfilled. AR/VR/“XR” do none of these.

It’s more exciting — and perhaps even more challenging — to think about technologies that are subtle, that fade into the background and gently nudge us: to connect with friends and family, to stand up, to take a moment to breathe. Imagine giving everyone a magical secretary that knows when to bother you about something and when to avoid interrupting, that can suggest delightful encounters, activities, and exercises, that helps you learn about new subject areas, and that helps you reflect on the quality of your decision-making through guided introspection. A world of true “user agents” that can help us become the people we want to be, and live happy and fulfilled lives. This is a direction where technology can nudge us toward self-realization and utopia. This is where I hope we can focus. Not headsets and solitude.

Speaking & Reading

Humans are really good at talking. Most people can talk at about 120 words per minute. Using a computer keyboard, most folks average typing 50, on a phone that’s more like 30, and you’re really in a pickle if you’re trying to tap out a message on a smart watch. Speech can be fast everywhere.

Conversely, your eyes funnel enormous quantities of data to your brain, much more than your ears. Reading is can be much faster than listening. It’s possible to “skim” text to quickly understand if a given message is urgent or casual in a way that’s just not possible with audio alone.

But just reading can often drop critical signal; sometimes it’s important to hear how someone said something. Were they being sarcastic? Were they making a joke? Were they crying while they said it? Or angry? These nuances don’t show up in text but can make a world of difference in how a message is interpreted. Preserving both speech audio and a good transcript is important for efficient communication.

Time & Intimacy

Another angle is asynchronous versus synchronous; busy folks don’t have time for very many phone calls — they need to be scheduled. If you just call someone out of the blue you may be interrupting their flow, busting in on a meeting, or barging in on a family dinner. So it’s more polite to just text instead, or passively read posts and status updates from the people we care about. But telemetry on the people is not the same thing as being close to them. And text messages don’t engender the same sense of connection and intimacy as a phone call or an in-person visit. What if we could have the intimacy of a phone call with the convenience of a text message?

Voice messaging is already a function of many popular instant messaging apps; and in many emerging markets it is common to share such messages. But a consistent point of feedback is that while it’s delightfully easy to compose them, it’s a pain to receive since you need to listen to the whole message just to understand its content. One user we talked to was in several Indian WhatsApp groups where many people were posting voice messages — just keeping up with the content in each group proved a substantial time commitment. People need to be able to scan the text, too! But no current instant messenger includes this critical use case.

Our team previously shipped an experiment called “BLRB” that let you publish blurbs of speech to the web as both text and audio. While novel, it only saw modest use — and when people did use it, they did so predominantly to voice journal or send personal messages to others. BLRB wasn’t built for either of these things, so we paused work on it and the team started chewing on what it would look like to build a product with messaging and journaling as first-class citizens.

Better Transcription

Most folks with an iPhone or iPad have probably used the built-in transcription provided by Apple, dubbed Siri. It was totally state of the art in 2010 but has languished as the industry has evolved past it. Transcriptions need many fix-ups and don’t even infer punctuation. Folks are used to transcription that doesn’t work very well, and don’t realize that things like inferred punctuation are even possible.

That’s why one of my favorite parts of introducing folks to Howdy is seeing their reaction to their first recording. “Wow, it gets the punctuation! I don’t need to dictate periods and commas!” “OMG it recognized everything I said!” People deserve to have apps that actually understand them — and that help them be understood by others.

Conclusion

Anna, Dima, and I have spent the last few months building this new experience, that puts the voice first in connecting people with each other. We’ve been using it to chat with each other regularly for the last few weeks and find it adds a touch of the personal that we don’t get on Slack and also lets us efficiently share longer and more complex thoughts with each other. We hope you’ll try out the app and give us your feedback! Email us at team@howdy.chat with your thoughts; or if you run into a bug, just shake your phone to report the issue to us.

I had the pleasure of working at Facebook many years ago. It was a very exciting time, where smart engineers from around the world got a chance to work shoulder-to-shoulder on problems that impacted a large and growing fraction of the human population. A lot of the excitement was due to an earnest belief that in helping people share with each other we would help people feel better connected to each other and that this in turn would lead to a better world. Making the world open and connected was explicitly our goal, and it seemed great.

My focus for a chunk of my stay was on understanding how to get more people online. We did a lot of work to understand the true reasons why people don’t connect, which turned out in most places to be not as simple as either a lack of affordability or a lack of access. We made the world’s first map of the disconnected and explored a hilarious diversity of connectivity technologies from drones to lasers to satellites. It was heady stuff, and then I got poached by Google to help them build a new research and development team to also explore getting people online, a mission which I was determined to assist.

Looking back, we had these naive beliefs around a connected platform being good because they were in part true. Facebook let you take people who were otherwise going to be extremely distant acquaintances, like that cool person you sat next to on a plane flight 8 years ago, and made them a “good acquaintance? who knows what’s going on with your life at a superficial level. You can continue to interact with them when otherwise both of you would have moved on. These kinds of weak connections can prove very powerful in helping people to network and in getting a wide range of inputs, fun questions, recommendations, and experiences. It makes things feel “social”. It yields the effect of making one feel vaguely popular, to have a lot more good acquaintances than otherwise your life might afford.

But the inverse of this was (by us) less anticipated — that by having a uniform platform where sharing and interaction leveled the playing field between intimate friend and distant stranger, we would not only pull the stranger in closer, but we would push our friends away. Now instead of having a handful of intimate friends and a handful of acquaintances, we end up with no friends and an impossibly large army of acquaintances.

In having the medium be permanent instead of ephemeral we find ourselves self-censoring for the judges of future history. And with a much larger pool of people there were more things that might be misinterpreted or generate drama. So the things that we shared tended to be positive and glib. But meaningful human connection is built from vulnerable and authentic sharing. Such “self-disclosure” has been well-researched to promote bonding between individuals, even complete strangers. Online platforms that share speech with a large number of people and preserve it over a long time are not welcoming places for all but the boldest to share their darkest secrets and fears. The result: a feed full of people you vaguely know, mostly posting nice things but not vulnerable things, leaves us feeling all the less like we can share our own struggles — and that having such struggles makes us unusual amongst our peers. This leads to alienation and further lack of deep sharing.

There’s nothing deliberately malevolent about these sharing platforms; these are accidental artifacts of how they were built; whole movies have been filmed about the impacts of optimising for engagement, which I won’t rehash here. But it’s the component of willingness to share vulnerably that I chew on, because a lot of what’s going on is our own miscalibrated expectations about the degree to which others want to hear what’s going on in our lives. And the surprising, delightful truth is that people do want to hear. But what’s needed is reciprocity and exchange: if we bare ourselves but are met with a “like” or a “hug emoji” then this does not create intimacy.

We have seen an exploration of the ephemeral with SnapChat and disappearing messages on chat platforms. We’ve seen a shift from broadcast behaviors to 1:1 or smaller and contained groups. I would expect to see more of the same, with new platforms offering ways of exploring intimate sharing and encouraging mutual vulnerability and authenticity. And I think this is a fine goal for what comes next: not just finding new acquaintances or getting reminded of birthdays, but building deep and enduring new friendships.

We really appreciate your trying out BLRB! This was an experiment that we launched in the Fall of 2020 to see if people wanted to publish ad-hoc transcribed voice to the web. As far as we know, nobody else had built a service like this before. We helped push the state of the art in mobile web audio recording and worked with the teams at Apple and Google to get their operating systems and browsers fully supporting our use case. We’re deeply appreciative of the feedback we got from all of you. It helped us refine our beliefs and hypotheses around voice.

BLRB, sadly, did not take off; after some early excitement with a viral share or two usage went to nil and stayed that way for most of 2021. Nobody outside of the team recorded a single BLRB in Q4. Equally telling, we as the product team didn’t find ourselves using it very much or excited to use it. So we thought about why.

Here’s what we found: speech is very intimate and raw — if we’re listening to a stranger’s monologue we expect something polished like a podcast. We’re shy to share the unscripted and bumbling versions of ourselves with the world. Most people don’t self-identify as podcasters!

But where people did use BLRB was to share that raw and unfiltered version of themselves with a friend or even as a voice journal; to help process their thoughts or to share authentically with others. In my own usage of the product, the majority of my BLRBs were actually private messages to others, even though technically the BLRBs were public.

So it was based on these insights and what we’ve learned from building the service (and Medcorder) that we’ve decided to work on an app called Talkie (TryTalkie.com) for authentic voice sharing to help you better connect with the people you love — you can send transcribed asynchronous voice messages to them, which are faster than a text message but more convenient than a phone call. We’re already using it to better connect with friends and family.

We’d love for you to add your email to TryTalkie.com and we’ll let you know as soon as a public build is available (iOS in Feb/March and Android a bit later in the year).

Thank you again for your thoughts and inputs and from our family to yours we wish you a very happy 2022.

Yours,
David & The BLRB Team

PS: We recorded a video demo of the BLRB service for posterity using Loom: https://www.loom.com/share/a4ce1748d3d341c0aed524bc3c9270e4

Today Sonic.net installed a 10 gigabit synchronous, unmetered XGS-PON fiber optic link to my home in Redwood City California. The connection will cost me $40/month. My only gripe is that I can’t get static IPs, too! I spent the earlier part of the week wiring my home with 10G capable switches and getting a 10G copper Thunderbolt 3 interface for my notebook — which means that for the first time in my life I’ve got a direct hard line 10 gigabit connection to the internet backbone.

What does that mean for Internet experiences? What comes next?

LAN Links

When I was still in college way back in 1999, gigabit copper ethernet was standardized as 1000BASE-T (802.3ab); it was 10x faster than its predecessor 100BASE-T, used the same wiring configuration (four twisted pairs, e.g. CAT-5 / CAT-5 cabling) and connector types (RJ45), and very quickly became a commodity connector for servers, desktops, and laptops.

Over the following two decades, datacenter fabrics continued to innovate on network connections; moving onward to 10gbps SFP+ through to now 800 gbps modules (QSFP-DD). This means that modern datacenter network connectors are literally almost a thousand times faster than the network links experienced at home.

Meanwhile, several efforts to improve consumer LAN links have slowly moved forward, resulting in the 2.5G and 5G 802.3bz standard emerging out of the merger of the MGBASE-T and NGBASE-T alliances, launching in 2016 but with only tepid adoption to date. Much of the motivation was running faster speeds over old CAT-5 cable but CAT-6/6e is now widely deployed. Another motivation was the ability to run power over Ethernet (which at the time hadn’t been incorporated into the 10GBASE-T standard) but in 2018 with the passing of 802.3bt, this issue was resolved. So today it’s unclear whether it will make sense to have deployments of 2.5/5G gear or whether it will be about the same complexity and price point to just go directly to 10G. Personally, I’m not sure I understand the draw behind 2.5/5G at this point.

Plugging gear into multi-gig links is the next step; Thunderbolt-to-10G connectors are ~$120 and 10G is now baked into high end devices; for instance it’s a $100 additional option on the M1 Mac Mini. We can expect pro devices to start incorporating 10G connectors next year.

Stepping back — why has consumer LAN connectivity basically languished for two decades while datacenter connectivity has grown by three orders of magnitude? Two factors have gated consumer adoption of multi-gig links: WAN and WiFi speeds.

The WAN Bottleneck

In the days of residential dialup, local connections were much faster than Internet connections (56kbps vs 100mbps in the mid 90’s; literally about a thousand times faster). This led to experiences like gaming that only made sense over a local network and the development of “LAN parties” where people would bring their desktops physically together to plug into the same network to play games, exchange music and video, etc. Over the following ~15 years, Internet connections rapidly increased in speed. In the US, gigabit fiber (GPON) & cable (DOCSIS 3.1) deployments started around 2005 and accelerated with the introduction of Google Fiber. As internet connections got faster people came to rely more on services outside their local network for backup, entertainment, and sharing; the purpose of the local network became mostly to connect devices with the Internet, moreso than connecting local devices together. Consequently, improving the speed of the local network beyond the speed of the Internet connection wouldn’t yield meaningful benefit. Why get a ten gigabit home network when all you’re ever going to do on it is connect to Internet services over a one gigabit link? Nothing will be any faster for you.

So in order for consumers to be interested in moving their LAN links to >1gbps, they really will need to have access to >1gbps WAN links. It’s only in 2021 that we’re finally starting to see meaningful deployments of >1gbps speeds in the US. While Comcast announced a 2gbps offering in 2015, it only had a handful of deployments until the last year or two — and it was only in the last 12 months or so that they increased the speed to 3gbps. Earlier this year Google Fiber lit up its first 2gbps homes in Kansas City. GCI has begun its 2gbps deployments in Alaska. Altice has announced it will start rolling out 10gbps deployments in 2022 — and Sonic has begun its 10 gig XGS-PON deployments, including at my house.

So multi-gig ISPs are finally here. Hooray!

I decided to plot out how my home connectivity speeds had changed over time since I moved into a shared nerd house in San Jose in 2001 and we splurged to get a $300/mo 3mbps SDSL line. We ended up installing a 19” server rack in a closet and hosting a bunch of interesting services. 20 years later I have a connection 3,300x faster that is 7x cheaper. That’s pretty astonishing.

The WiFi Bottleneck

Most devices in a home connect to the network over WiFi. In fact, most residential deployments are WiFi-only, with very few if any devices plugged into Ethernet (despite that fact that hardlining your connectivity is a very good idea: lower latency, less jitter, faster speeds, and more airtime for your other devices). WiFi as-implemented makes reaching speeds of over 1gbps of “goodput” (actual bits transported) improbable with typical consumer devices. Wi-Fi 6 helps, but IMO the real gamechanger is Wi-Fi 6E which opens up a vast amount of new spectrum in the 6GHz band (confusingly enough this has nothing to do with the “6” in “Wi-Fi 6”).

Regular Wi-Fi works in two separate spectrum bands: 2.4Ghz (with a relatively narrow slice of 60MHz of spectrum that’s incredibly crowded, shared with everything from microwaves to Bluetooth) and 5GHz, where there’s a lot more spectrum (~735MHz worth in the US!) but with a few drawbacks. First, most of the spectrum in the US is subject to Dynamic Frequency Selection — the US military only allows use of certain bands (most of the 5GHz WiFi spectrum) meaning that both access point and clients need to be ready to vacate the band very quickly after detecting a military radar. That creates extra burden for both in both monitoring the current channel as well as monitoring their “Plan B” channel for where it would be safe to switch were radar to be detected. This is complex and drains batteries so a lot of implementations avoid or disadvantage DFS band APs in complex ways (e.g. roaming may end up partially broken, discovery of APs in DFS bands can take substantially longer, etc). So there’s less “easily usable” 5GHz spectrum than one might think; if you’re excited about “wave 2 802.11ac clients” that can use 80MHz channels, there are only three outside the DFS band in 5Ghz. And only one of those can be used at high power.

Now those of you who have looked at access points know that the marketing lingo on them insists that they are much faster than 1gbps (NetGear is advertising their “10.8Gbps” Nighthawk router), but the marketing speeds are complete nonsense — not just from NetGear but from all the vendors. They add up the theoretical maximum PHY rates for each of the bands to get a non-sensical sum capacity. To make a bad analogy this is a little like when asked how fast your car can go adding up how fast your car can drive in first gear and adding it to how fast your car can drive in second gear and so forth. “300 miles an hour!” The assumption is that you as a consumer are too stupid to understand that the number in question is a lie and that you’ll just lean hard on bigger numbers being better. The reality is that with a 2x2 80MHz client you’re very unlikely to break 1gbps in the 5GHz band.

The 6 GHz band enabled by Wi-Fi 6E brings an additional whopping 1,200MHz of spectrum to the party; this should enable practical use of 160MHz channel clients (if you really want to geek out you can check out the full MCS table); and it’s only at these extremely wide channels that we can expect a 2x2 client to be able to actually achieve >1gbps of goodput with WiFi. While there were rumors the iPhone 13 was going to include 6E support, it looks like that didn’t happen, so we’ll need to wait for at least the iPhone 14 on that front; we’re about a year or two out from a critical mass of 6E-capable clients.

Even with Wi-Fi 6E we’ll only just be nudging above a gigabit for most clients; even “premium” clients like the 2021 MacBook Pro are just using two spatial antennas (somewhat surprisingly this was a regression from the 2017 MacBook Pro which had three), so even in a year or two it will be pretty unlikely for a typical client connecting to even a high-end WiFi access point in typical conditions to get much over a gigabit.

This means that most of the benefit of the multi-gig deployments is going to come from either a very large number of very busy clients, or from clients that bother to plug in physically to the network, which will require a bit of a mentality shift for folks.

Deployment

In order to cross-connect the XGS-PON device to my UniFi Dream Machine Pro at 10gbps I needed to plug the 10GBase-T into my SFP+ port which required buying an additional adapter. Sadly, the UDMP doesn’t support load-balancing multi-WAN so my hilariously overwrought plan to use both AT&T gigabit and Sonic multi-gig fiber together isn’t doable with this hardware setup. So the moment of truth involved unceremoniously unplugging AT&T to cause the UDMP to failover to my multi-gig WAN2.

Benchmarking with SpeedTest and Fast.com is showing closer to 3–6gbps down and 3–6gbps up depending on the phase of the moon and the weather — it’s not a full 10gbps yet, but it’s quite a chunk faster than 1gbps. At these speeds, a lot of factors end up coming into play that start potentially bottlenecking performance: client hardware and operating system, choice of NICs/switches/cables used to connect, and ISP health.

One surprise was that my WiFi end-to-end benchmarks ended up meaningfully faster; peak WiFi6 speeds from my iPhone 13 Pro Max had previously capped out around 450mbps down & up; now I’m seeing as much as 700mbps both ways! I don’t have a great explanation for why improved backhaul would have made this level of difference in realized WiFi speeds.

Conclusions

We’re on the cusp of seeing multi-gig Ethernet finally deployed in US homes after two decades of plateauing at 1gbps. 10G ethernet technology is mature and ready for deployment on typical residential CAT-6 cabling. While WiFi will get much faster with 6E, consumers will need to get used to plugging in devices for which they want top speeds, which is a change in behavior from today, where WiFi was about the same experienced speed as plugging in. People doing construction on homes should consider pulling CAT-6A cable quality to support 10G copper runs; getting a nice cable only costs a small amount more than lower quality runs. To fully future proof, consider installing conduit to each room in the house to make pulling new runs easy as new connection standards are released; it’s not hard to imagine homes in just a few years switching to 25G or even 50G optical networking, which will mean pulling new cables through the house.

But “so what?” What changes in terms of people’s experiences when they have 10gbps connectivity from their devices to the Internet? The short answer is: I don’t know yet!

I do know this: at 10gbps connectivity, the Internet can be considered to be on a “data bus” with your local computer. Nearly any device that can plug in physically via USB-C can be virtually plugged into your computer. Storage can be completely virtualized; limitless in size and almost as quickly retrieved as if it was on a hard drive attached to your computer. (Yes, NFS has been around nearly 40 years, but now we’re finally in a position to really treat remote volumes as local without much of a haircut!) As a developer the boundaries between interacting with my processes locally versus those running in the cloud become very blurry; I can burst container instantiations into the cloud and cloud processes can burst data to be processed locally. My device becomes a “peer node” on the Internet instead of having to just participate by having deeply digested information spoon-fed to it.

The 20th century model of life can be roughly encapsulated as being in three acts: in the first act, one is educated up to the limit of one’s mental capacity. Then, once full of knowledge, you are sent out into the world to work for as many years as you can. Finally at the end of your career you retire, have some fun, travel a bit, then die. We can call these stages: Learn, Work, Play and they are mostly accomplished serially. This worked acceptably well for a world in which the exercise of one’s craft at the end of life roughly resembled that at the beginning and in which retirement was likely short since life expectancies were only modestly longer than retirement age.

But in the 21st century, this model is breaking down. First, as we shift to more of a knowledge economy, the number of years required to obtain sufficient knowledge to be useful is continuing to grow. The percent of Americans attending college has increased, as has postgraduate education. Some don’t begin their working lives until their early thirties! This extended academic interval accompanied by the destruction of teenage jobs and apprenticeships mean that young people don’t get a chance to perform useful work until quite late in their lives. I hypothesize that this is a driver of disengagement from school — why spend hours laboring over an essay that will only ever be read by one person — skimmed by a TA to have a grade stamped on it — instead of composing a silly TikTok that could go to hundreds of thousands of people? Why is one “real” work and the other is “play”? Teenagers have the mental and creative capacities to be meaningful contributors to our society and culture; we should stop being surprised by this but instead seek to incorporate it into learning; to have their work Matter as early as possible and to use it as a teachable moment.

Secondly, the 21st century has jobs and fields that are evolving quickly. Many key disciplines today don’t have traditional college degrees that cleanly map to their job functions. And while many jobs purport to have “on the job training”, formal educational programs onsite are rare. At Google, where I had the pleasure of working for over four years as a senior product manager, there was not a single formal class for existing product managers to learn more about their craft and advance; and this is from one of the world’s leaders in product management, with a talented and well-funded corporate education department! Cultural expectations around “adult learning” are dismissive and remedial — community college and job retraining programs are often viewed as a rare and surprising move to change career tracks when one has hit a dead end instead of an expected and regular part of adulthood. Ongoing formal instruction should be embraced for all adults and expected of all knowledge workers. This should include periodic paid time off to “level up” — such investments in people’s capability will likely see returns in spades.

Finally, we can’t forget the “fun”. Knowledge work in particular shows rapidly diminishing returns for hours worked beyond 50 a week and several experiments reducing the workweek to four days (or even three) seem to be showing positive results. As “rote” deskwork is increasingly automated, the creative cognitive burden of the typical job is likely to increase — demanding inventiveness and problem-solving that require fresh thinking instead of mere “knowledge grinding” (as is typified by late 20th century work such as data entry). A working environment that is enjoyable is going to be more sustainable and can therefore retain employees longer, also allowing employees to continue meaningful work well past traditional retirement age. (Classic “retirement” is mostly a Bad Idea anyhow and leads to increased mortality and decreased satisfaction.)

Returning our focus to education: what is its point? Ultimately, it is to produce the citizens that we want: knowledgeable, wise, clever, creative, coöperative, resilient, and hard-working. People who are able to perceive problems, frame them usefully against past efforts, hypothesize meaningful solutions, test their hypotheses critically, work across a team to organize and accomplish required tasks, and evaluate outcomes. We want to maximize the personal potential of each student.

We then have to ask whether our current educational system is efficient and effective at achieving this goal and what might be already known about potential improvements to the same.

So let’s start with an overview on the learning experience of the typical American student:

• Fixed-duration classes: a class is defined as the teaching of a subject for a certain number of in-room sessions per week over a period of a number of weeks. At the end of the class, the instruction is deemed complete. Students have a final exam where their completeness and correctness of understanding is evaluated as a written test.
• Multiple classes per unit time: students are typically taking 3–6 different subjects, different in nature, contemporaneously.
• Group, lecture-based instruction: a qualified teacher orally presents material to a group of 20–30 students in a lecture hall or online, using multimedia aids such as a chalkboard, whiteboard, or prepared slides.
• Homework: students are assigned reading and exercises to perform on their own at home, typically several hours’ worth per day. These exercises are intended to solidify understanding from lecture as well as to grade progress. Little 1:1 instruction or assistance with homework is given (with the exception of TA sessions in college) — if students require extra help understanding or mastering the material they are expected to find such help on their own (from parents, tutors, etc).
• Instructor-driven grading: the instructor is also generally the person responsible for grading a student’s homework and exams. In subject matters where subjective scoring has a component, a teacher’s subjective impression of a student may influence grades given. Independent evaluators or examiners are rarely used; intervention by outside forces such as a principal, school board, or district, happen only in extreme, unusual, and Bad cases.
• Early start: instruction typically begins early in the morning, typically 7 or 8am.

We should feel almost embarrassed to write it out this way because almost every one of these has been shown to be a Bad Idea.

Mastery Learning

If a student with a poor understanding of one subject A (say, a ‘C+’ grade) goes on to take a further course B with A as a pre-requisite, they will likely struggle with B and similarly perform poorly. This leads to a chain of failures very difficult to remedy that frequently leads to loss of confidence and joy of learning; the student self-identifies as being a poor learner. Conversely, if the student is allowed to remain with the material until it has been mastered, then all students taking course B can be confident that they are ready and have mastered its prerequisites. Mastery learning has been shown to improve educational outcomes and learner confidence. Similarly, if a student happens to find a given course straightforward, allowing them to advance at a pace matching their understanding keeps them challenged and engaged. This is not an argument for “tracking” of “good” versus “bad” students; one subject may be easy for a given student but its successor challenging — and vice versa. Allowing each student to progress at their own pace can have the opposite effect from tracking by enabling those who took longer with earlier material to “catch up”, whereas almost no such opportunities exist in a tracked environment.

Focused Learning

Intensive periods of focused learning about a single subject lead to much more rapid advance in understanding of that subject than interweaving many different subjects. This is because there is substantial cognitive burden in performing a “context switch” from, say, Descartes to Asian History to Spanish to Calculus. While to some a blend of many different intellectually stimulating inputs coming contemporaneously can sound exciting, our system of instruction should be focused on that which is effective. In several domains outside the educational system we can directly observe the difference in learning rate between focused learning and interwoven learning — in aviation there are accelerated flight schools that a pilot can use to achieve their instrument or commercial license in two weeks (flying 6–8 hours a day) as opposed to six months of interwoven instruction (flying 4–6 hours a week). The military notably also uses such intensive training bootcamps to rapidly provide effective education.

Spaced Repetition

To have long-term retention and recall of critical knowledge, we know that regular use of the material is required; specifically using “spaced repetition” to return to material. Current curriculum design does not incorporate spaced repetition, instead focusing the learner on “cramming” for a final exam — this loads a tremendous amount of information into short-term memory for test-taking purposes to immediately subsequently discard it. If the purpose of taking a class is to have the material continue to be retained and accessible by the learner, it will be required to provide accommodation for periodic review and refreshing of the material over an extended period of time, such as a year.

Teach to Know

In many disciplines, a student is not considered to have an adequately rich understanding of a subject until they can teach it proficiently and answer questions from other students about the matter. For instance, the level of knowledge and clarity of explanation expected out of a certified flight instructor (CFI) by the FAA is much higher than that expected of a private pilot or even a commercial pilot. Somehow this obvious insight is mostly ignored in education; beyond some graduate students acting as teaching assistants, most students are not expected to or offered the opportunity to instruct. An expectation of instructing on material after mastering it presents both an opportunity for further mastery and a natural supply of tutors. Furthermore, the ability to teach — to share knowledge and observations with others — is a critical component of a knowledge economy. Therefore the fundamentals of instruction and pedagogy should be required for all graduates, not just those specializing in education. Many universities tout themselves as a place to “learn how to learn” but almost none provide a foundation in pedagogy to all their students.

Tutelage

One of the most shocking outputs for me in reading Bloom’s seminal 2 Sigma Problem paper from 1984 was the core finding that tutoring instead of classroom instruction caused a two-sigma effect; students that were tutored performed in the 98th percentile. This dramatic effect throws into question the whole premise and framing of traditional classroom instruction as being wildly ineffective. (Don’t blame the teachers! The medium itself just doesn’t work well.) While Bloom’s take on the paper was to assume that such instruction couldn’t be scaled and consequently to try and explore methods of ameliorating the downsides of classroom instruction, Bloom overlooked the obvious solution of turning students into teachers to scale tutelage.

Scaling Instruction

If tutelage is this dramatically effective at instruction, less overall time may be needed in order to achieve the same educational outcomes. Conversely, outcomes may be advanced substantially in the same fixed period of time achievable by traditional instruction. Similarly, tutelage of a range of students at varying levels below an instructor naturally provides the spaced repetition called for above. In flight instruction, a CFI may find themselves one day offering a demonstration flight to a person who has never taken the controls of an aircraft and the next day helping a pilot with 1,000 hours achieve their Multi-Engine Instructor license. This variety keeps things interesting and exercises a CFI’s working recall of subject matter basic and expert alike.

Age-Appropriate Instructional Hours

The evidence against early morning instruction for teenagers is crystal clear: it’s a bad idea [Kelley et al 2013]. Consensus extending up to the CDC & Department of Education is that instruction should begin around 10am to allow teens enough time to sleep and be ready to receive instruction. As we continue to learn more about the interplay between biology and capability we should adapt environments to maximize learning rates. Examinations and evaluations should similarly be given later in the day to accurately measure peak performance of the learner.

Oral Exams

A written test alone is not sufficient to establish depth of knowledge and ability to apply knowledge to practical problems. A combination of a written exam followed by an oral and practical exam (similar to what is done by the FAA) can allow an examiner (a separate individual from the instructor) to dive deep into the areas a student got wrong on a written test and establish whether or not mastery has been obtained. Figuring that a competent examiner could perform about 15 oral examinations a week in a given subject matter, one examiner could oversee mastery verification for north of 500 students a year, allowing the position to be highly leveraged and appropriately compensated. Instructors would “sign off” a student as having mastered material before their written and oral instruction, and a low pass rate by an examiner would reflect poorly on the instructors who had signed such pupils up for examination. The final goal of a student in a given discipline, then, should be to assist with instruction of other students who go on to have a high pass rate in their evaluations. These oral exams could be recorded to be randomly spot-checked by a district administrator to ensure consistent and fair practices and evaluation between students.

While the U.S. educational system doesn’t make pervasive use of oral examinations (with a notable exception for the PhD defense), its use is the primary method used to assess employability; the job interview is itself an oral examination of a candidate’s preparedness for a role. Such disparity between corporate and educational evaluation should be concerning, as students often come to their first interviews underprepared for such an encounter, having never had such an interaction in their years of schooling.

Approach

Putting all of this together, this suggests a system whereby even from a young age, we make materials available to students for mentored/tutored studying where most of their teaching assistants are older and more advanced students themselves. Seasoned instructors can supervise the tutelage and offer guidance to both the students and the teaching assistants alike. Students progress at their own pace and “complete” a course to the first order when they passed an oral examination and to a second order when they can consistently teach other students to master the material. Grades are not given; students are judged instead by the quantity of material they have demonstrably mastered and competently instructed on.

Toroidal Learning

Traditional instruction requires the student to accept on faith that the subject matter being taught will ultimately be useful to them after school — either directly or as fundamental for understanding later practical material. For those for whom the material is itself exciting that is sufficient, but many learners require context: “why is this helpful to know?” that is missing from instruction and curriculum design outside of trade schools. One approach akin to that used in Montessori school is to embrace that depth in one area within the context of existing in the world ultimately ends up relating a multitude of subjects; to visualize it, going very deep leads eventually to a broadening that encompasses everything — when you dig deep enough and come back around to your starting point from multiple directions that is the topology of a torus and consequently I call this model “toroidal learning”.

Let’s suppose that a student is interested in horse racing; not generally a subject taught in most schools. Going deep on horse racing can end up meaning coming to biology (the breeding of horses), physics (the mechanics and motion of horses), advertising & marketing (making people aware of your venue), sales and statistics (how you price odds and entry fees), government (where you can zone and develop new racetracks, permitting), construction, law, economics, and more. Since all subjects exist in the world, understanding the interplay between a subject and the world requires an understanding of the world to understand a subject.

Consequently, a self-driven curriculum design suggests a “tech tree” of interdependent modules of knowledge that may be smaller than traditional courses, allowing us to jettison the notion of fixed-length classes (something already mostly required from embracing mastery-based learning) and couple knowledge with a practical working area of interest to the student.

This model would require a mentor to assess a student’s interests and knowledge and assist them in mapping modules of knowledge applied to their interest. Output could be assessed by a metric appropriate to subject domain and could include apprenticeship and other real-world exposures.

The Real World

A student’s work should matter — ideally this happens naturally because the coursework they are performing has interesting and useful application outside of the classroom. A good example of “real world instruction” is the Facebook bootcamp, a six-week program for incoming technical hires that combines practical classroom instruction about the architecture of Facebook coupled with hands-on tasks resolving actual bugs with various portions of the Facebook codebase. The ability for a new hire to within a week or two submit changes to production for a few billion people confers a sense of agency in the company that can be exhilarating. Freedom to exhibit mastery in a subject should ideally be permitted to take different forms; a funny TikTok with a hundred views about a review of a classic book could be an appropriate submission, for instance. A poem or a wood carving of a key scene in a book could allow students with mastery of other areas to bring a creative take to the subject.

For youth, education could be interspersed with internships and apprenticeships at an earlier age, to give them exposure to real working environments and conditions. While acknowledging the complexity of introducing minors into a professional work environment, the upside potential of motivating students to achieve and giving them context for their learning could be outweighed by the risks and burden of mentorship and oversight. As large tech companies have discovered, an internship program not only allows for useful work to get done but also provides for a regular influx of qualified talent ready to hit the ground running at graduation. This results in larger starting salaries for graduates, well into the six figures for technical offers.

For adults, incorporating regular structured learning (not just informal learn-by-osmosis “on the job” training) and instruction into job roles should be considered as a core part of knowledge era employment. Carving out short windows, perhaps two weeks long, to engage in focused and intensive instruction could enable employees to “level-up” in key skills. Expecting senior contributors to periodically teach their craft could help provoke interest from qualified applicants.

Summary

Our current educational system is highly inefficient and ineffective; we’ve had the data to show this for decades. What we know about how people learn and how instruction can be effective has been systematically underapplied. But we can’t afford to squander people’s potential. What we need is a system to systematically apply an understanding of pedagogical research by making learning intense, personal, relevant, and lasting to maximize the excitement and impact of learning, and to continue this process throughout everyone’s life and career.

Thoughts? Comments? Please contact me at david@weekly.org. As a disclaimer, I’m underqualified to write this piece as I am not an educational researcher nor do I have the academic background to firmly ground my opinions in peer-reviewed research. Consequently, studies and data that either support or refute my hypotheses are warmly welcomed and will be incorporated into the piece.

I know the schtick is to “start with the small stuff and then build” but the purple iPhone and reskinning of the Podcast app were pretty underwhelming. There’s so much more Apple could be doing in the podcast production and consumption space. Boo.

AirTags: long-expected, they look delightful. A little overpriced as typical. I will probably get these. Great Find My + UWB integration and privacy considerations here. Very impressed they led with a “unwanted tag” use case addressed out of the gate. Apple is clearly and consistently pushing on magic *with* privacy. Does it make sense to put one in a car to find it in a lot? Stitch it into a wallet? There wasn’t any screen time spent explaining how charging works for these.

AppleTV 4K — (are we supposed to call it “The New AppleTV 4K”? Why is Apple so bad with names?). Academically nice to have a faster processor and HFR HDR is theoretically nice but how many TVs can support this? Are there bitrate issues with receivers/cables supporting? I wonder how many folks will buy these expecting improved visuals and end up disappointed. Apple even alluded to the fact that almost no content is available today in this format. If the processor is a lot faster, shouldn’t that enable new gaming experiences? Almost shocked there was not even a tip of the hat toward this — it makes me wonder if “gaming on AppleTV” has completely tanked as a use case to not even be worth mentioning. The iPhone color correction is a solid idea but not at all clear if this will work with projector systems. I’m probably not going to get this. Missed an opportunity to catch up to Chromecast / FireTV on form factor (bring it with you HDMI, which is honestly pretty magical). Looking at the details it’s an A12 Bionic which first came out in 2018. Ah well, guess they don’t have enough downbinned M1’s, or they wanted to find a cheap dumping ground for processors that are three generations old. Underwhelmed.

iMac: they keep toggling between colors and not. I mean, hey, orange is fun. But why on earth did they keep the *chin*? Was really expecting a near bevelless design. Especially if you’re leading with a tag line like “make the computer disappear”. At least they finally put something better than a crappy 720p camera in but it’s more than a little shameless that they are patting themselves on the back for including a 1080p module half a decade after their competitors started doing it. There were touchpoints on image processing but it wasn’t clear how configurable that would be. (side note: OMG NAVPREET IS BRINGING GREAT ENERGY — I love this guy) it’s not hard to imagine how they could have been more radical with the speaker design, (like turning the rear face into a bass emissive element or using top-firing tweeters for Atmos style surround-bounce) but six speakers is interesting — will have to hear it to see if it’s impressive. On the M1 architecture I have no idea why you want to highlight that e.g. Twitter’s app is now M1 optimized. What the heck are you tweeting that takes a state of the art computing system?? The pitch to play a video game while on a Zoom call was surprisingly tasteless/tactless. C’mon Apple, is this the future you want?

A little sad we won’t see M1X today. That’s probably going to be for the Fall in the iMac Pro 2021 and 16" MacBook Pro.

They now have a magnetic power connector for their desktop but took it away from their laptop. Say that out loud. This made sense to someone?? The Ethernet plug in the power brick is cute but it’s not multi-gig, and they don’t comment on WiFi 6E capabilities for wireless so I guess they are shipping with dated LAN connectivity technology too. Boo. And oh god *8GB RAM* on a $1500 desktop computer being sold in 2021??? APPLE!!

Ok they had too much fun with the Tim Cook Mission Impossible sequence. Ditto the “oops I was on mute” Zoom Cameo.

12" iPad Pro: 1600 nits? Jeebus that is delightfully searing. Why couldn’t they put that display on the iMac (which has a comparatively puny 500 nits)? The display looks amazing. And 2TB storage as an option is great. They will sell a bunch of these. Though I guess they aren’t launching a new Pencil? Real-life LIDAR implementations are still pretty lacking, though — wonder if this sensor will end up going the way of Force Touch.

During the presentation there were several hat tips to the sophisticated image processing that is now being done on board, but very little exposed in terms of how this could be played with or modified for creative control around personal appearance in video conferencing.

Based on what we saw here today, my hunch would be that in the fall we will see a 16 inch MacBook Pro and iMac Pro both with the gorgeous XDR screen and an M1X chip. Tragically it is looking like even the M1X might have a 16 GB RAM cap so we might be well into 2022 before we see an M-class chipset that can support reasonable levels of RAM (>=32GB). BOOOO.

Given the strong position Apple has taken on 5G (even including giving airtime to 5G UWB which is almost entirely academic in realized deployment), it seems all the more glaring that they aren’t similarly offering either WiFi 6E or multi-gig wired Ethernet in their connectivity options yet.

On the hard-core technology front I would imagine that Apple if they wanted to could do some pretty incredible things with Optane to make it feel like a machine had, for instance, a quarter TB of RAM. Imagine real-time scrubbing through hours of 8K video. One of the reasons why we haven’t seen more interesting developments here is that only a handful of players like Apple have the full stack control and capability required to deploy and realize the power of a new class of persistent storage.

I am also looking for leadership from companies like Apple to solve the “eye contact problem” — where a key missing ingredient of multiparty discussion is the ability to look each other in the eye. Some of this could be mechanical, such as putting cameras behind the screen, and some of this could involve using machine learning to subtly alter pupil alignment, as NVidia has been demonstrating with Maxine.

No mention of wireless charging of *anything* today. Once burned, twice shy?

Is it weird that I feel it was a miss for the iMac to not be compatible with Apple Pencil? There’s already the magic keyboard for iPad to make it in close to the form factor of a smal iMac, but the converse is not true. Would it really have been that much harder to add Pencil support and a little magnetic mount for attaching Pencil to the side?

Overall today’s announcements were decent but not mind-bending. I’m looking forward to hearing the new iMac and seeing the iPad Pro display. I don’t think the Apple TV upgrade is worth it for people who already have one. The AirTags will probably be pretty popular. I am a little frustrated that it will probably not be until next year that I can buy an Apple machine with decent RAM.

With Apple in a mode of executing well but iteratively, it does leave open the opportunity for some other players to step in and offer radical new full stack solutions. Microsoft is playing earnestly in this space with Surface, for instance. Google on the other hand has only moderately been playing in the first party consumer computing space (having abandoned tablets after the 2013 Nexus 7 and only investing lightly in innovations with Pixelbook). A refreshed commitment to driving state of the art with machine learning integrated across both local client and cloud could play well to Google’s strengths. And while Amazon’s prior efforts with their “Fire” line were mostly catastrophically bad, I would be delighted to see a continued exploration around voice-centric computing and leaning on the cloud for beefy coprocessing to save the local client effort.

That brings me to my last thought: Where is Siri? In that whole hour-long presentation I barely saw a mention. Siri has been languishing in Apple’s hands since the 2011 acquisition — it would sadden me if Apple’s answer was to give up and put Siri on the sidelines (as they literally have done with the new AppleTV remote) instead of buckling up to improve it to deliver true voice computing à la “Her”.

May your children be filthy,

Dirt on their feet,

Brambles in hair

May they be tired

Bags under eyes

From reading late under blankets

May your house be messy

Filled with screams and roars

Of laughter and life

May your days be filled with light violence

Of wrestling and throwing and spars

May your meetings be too short,

So much said but unsaid

That more must wait

May your friendships linger long in silence

For all that doesn’t need to be said

Quiet companionship an answer to the most important questions

May you be bored atimes

Wandering lost in the wood

Agape at nature’s bounty

May you be lost in life

Unsure what comes next

Or what your narrative is

Seeking to be useful and kind

May your path thus be uncertain

Nothing preordained

Keeping everyone guessing

Even yourself

Until the last

May your life be filled with confusion

And love

(For that is love!)

And side quests that are the main quests

And extras who become protagonists

And the delight of connection to others

Betwixt others

Introducing

Building

Forming

Hoping

The willingness to be

(In spite of all weirdness)

That which you are:

Unembarrassed

Unadorned

Except by your own peculiar quirks

Which seemed a distraction,

Which seemed a distress,

But now are the main attraction!

Nobody else has got the scars that YOU bear

And Oh! Does that make you unique and delightful

Your moment is brief

Now go shine!

The universe will never be the same after.

This morning I was reflecting on the fact that it has been some 20 years since my primary job was computer programming. It’s quite possible that I would struggle — or even fail — at a number of the technical interviews that I passed in the past. This is not a sign of my dotage (yet!) but rather a reflection that my focus and skills development has been in other areas. I haven’t maintained my proficiency entirely.

In aviation, we have this notion of being “recent”: you need to have practiced a set of skills regularly, or if you’ve fallen off the wagon and haven’t flown lately you need to have those skills examined to verify proficiency. As a private pilot, once every other year, you need to go up with an instructor for an hour to get a refresher. All of this is codified in law and regulation. The FAA wants to be confident the pilots in the skies know what they’re doing and are safe to fly. On top of that, at nearly every level you are expected to be continually learning and improving your skill set, and there are a wide range of tools and services to help you with that. all of that makes a lot of sense and is completely unsurprising.

More surprising, then, is to turn one’s attention to other areas where a high degree of proficiency and continued development are critically important, and the degree to which there are not analogous expectations and services as what we see for pilots.

I suspect that this is in part due to a 20th-century worldview: where we believe that education is a phase in life. We receive as much education as we can tolerate or afford, then proceed onto the next chapter of our life of putting that education to work in a job, then if things went well we can retire comfortably and die at peace and satisfied. But this model no longer holds. The increasing amount of education required to have a meaningful impact in the world means that the education phase pushes out farther and farther, leading to teenage frustration and angst about being unable to meaningfully interact with or impact the world since the job phase has not yet arrived. (And a good deal of stamping the natural fun out of the young educational years since “this isn’t meant to be fun, it’s meant to be learning!”) There are similar tensions in mid-life around deferring fun to one’s retiring years — and then finally the despair of disconnectedness that can arise from a detached retirement at the end of life. Our 20th century template for a good life is not a good one.

What we need is a template for a life that can combine continuous education, impact and engagement with the real world, and fun. It’s a theme I’ve touched on a few times, but I think it’s important for our societies health that we start to think through a lot of the assumptions that we bake into everything from our schooling to our work policies to our retirement plans.

And to wit, it is incumbent on every one of us to continuously be finding ways to learn. To refine our craft, to share our observations and mistakes, to seek out expert counsel in books and mentorship and peers on subjects that are important to us. That includes taking ourselves seriously as teachers, not only for others but for ourselves as well. Asking ourselves the questions, contextualized by actions we can perform: what’s going well? What’s not going well? What are the areas where I need to learn more and how can I best learn those things? If I’m not sure of my answers to any of these, are there resources I can lean on to help provide guidance? What are areas where I have learned things that could potentially help others? What could I do to effectively share that knowledge with others?

Finally, to close on the topic with which I started: what are areas in which you would like to maintain proficiency, and what are your plans to do that? Have you set aside the time and the mentorship? Could you go through a book or take an online course? Are there tools that will help you stay sharp at your craft? It was an obvious but important realization for me that even Olympic athletes have coaches. Especially Olympic athlete have coaches.

At whatever level of performance you find yourself, there are resources out there to help you excel. Find them. Level up. Go forth and kick ass.

Learning and Proficiency was originally published in Age of Awareness on Medium, where people are continuing the conversation by highlighting and responding to this story.