#84 | Full Context Companies, Frozen, Florence, Kindness and Misfortune

My irregular newsletter on work and play. Unsubscribe at the bottom. LinkedIn

And just like that it’s been over 3 months! Quite a bit happened.

MENU

1. WORK: Matchups, Deep Tech Summit
2. HOW I AI: Full Context, Learning Rate, AI Personas
3. CULTURE: Sultans, Beef, Bref, Larry, Frozen, Florence
4. THOUGHTS: Being kind, Misfortune

1. WORK

Events

• The SOSV VC/Founder matchmaking event focused on Physical AI ended with 933 participants, including 435 investors. Another focused on Therapeutics had 574 participants.

• Those free virtual events are becoming very popular, and remain an excellent source of deal flow for us, and of investors for our portfolio. The next one is focused on Deep Tech (November). I’m also preparing a “deep tech cofounders search” matching event for November. Pre-register to either or both here.

• The SOSV Deep Tech Summit - a new and improved version of our long-standing Climate Tech Summit - is coming in November. It will cover 3 themes: Climate, Health and AI for the Physical World. We have announced the first dozen speakers, including Jennifer Doudna (CRISPR Nobel), Steve Blank (Stanford / Lean Startup), David Kirtley (Helion / fusion) and more.

• I moderated a panel at the always great Hello Tomorrow in Amsterdam on “Why Do The Best Deep Technologies Still Fail?”. I was glad to have two founders of “failed” companies (one is rebooting after bankruptcy) and Michael Jackson on this panel: he’s one of the most interesting and humorous VC voices on LinkedIn with an astute cross-border perspective (and a quarter million followers). Most VCs and founders don’t talk straight because they care about their current job, company, investors or future prospects so this was a rather unusually frank discussion.

2. HOW I AI

Projects

• Over the past year, my understanding of AI has changed dramatically. I started by using it to write small scripts and automate personal tasks. After building more than 200 AI projects, I think the technology itself is no longer the main challenge. The hard part is designing reliable systems with end users (co-design), integrating them into organizations (integration), and helping people actually use them (onboarding / customer success).

• I still don’t see myself as a “software engineer” but I’m definitely a builder. My latest project is about 8,000 lines of JavaScript running as a Google Apps Script app. And I’ve never written a line of JavaScript.

• On the deployment aspect of projects, I also found that the technical prototype is not even half the battle. The real blockers are data permissions, co-design with a motivated and tolerant end user, onboarding and of course… data quality.

• Builders need to become product managers and also “customer success manager” — making sure end users see how to take advantage fully of the tools. The handover process should be 4 steps, and most companies stop at 1 or 2 (“deploy and forget” or “demo and forget”):

1. Show examples of outcomes, and one example end-to-end

2. Do one example together with the user

3. Let the user do an example, with support

4. Handover

The Full Context Company

• In the last letter, I mentioned a scale for personal AI fluency.

• My current view is that most organizations actually need to level up their “organizational AI fluency”, from using chatbots to shared skills relying on common context.

• This context is built from their various sources of data, from financials to emails, slack messages and meeting transcripts. Data dies in silos.

• “Full context companies” will use AI tools to guide or execute tasks fully aligned with the company’s roadmap and values.

• It is easier for new companies to build that capability from day 1 than for existing companies, which have a lot of legacy systems, processes, behaviors and… people.

• New versions of existing companies could likely be operated with a fraction of the staff, and several friends have started new companies with just a few team members, with no plans to expand the team.

Learning Rate

• A company’s biggest competitive advantage is often not intelligence, capital or even talent. It’s learning rate. Learning comes from experiments. Experiments multiply when friction and cost are low.

• The challenge is thus: how to reduce friction and learn fast not only as a collection of individuals, but also multiply the impact of learning by sharing internally effectively.

• See this excellent article by Steve Blank on the topic. How you operate with AI and how fast your company learns will make all the difference.

AI Personas

What is your view on AI:

• AI Skeptic > You tried a few times and were underwhelmed by the results. Skeptics generally underestimate the progress of AI, for lack of direct applications to their work.

• AI Oracle > You believe that you can just throw data at an AI and it will make sense of it miraculously. The reality is - as always - that while it can work on occasion, it generally fails due to (1) erratic results (2) performance as well as (3) “editability”. Not everything needs AI.

• AI Pragmatist > The AI Pragmatist knows when AI shines, and when it should stay out of the way. While I use AI for coding and never write a line of code by hand, my experience has been that if you want reliable results and fast performance in the system you build, you actually want to put in as little AI as possible, and instead rely on SQL-like queries and “clean” workflows. The most useful tools I am building for SOSV use zero or minimal AI for the end result.

Courses

• My vibe coding Udemy course is now FREE.

3. CULTURE

SHOWS

Beef 2***
The first season with Ali Wong was pretty funny and the second started very dark, which almost made me abandon it. Eventually it was a good show, with great acting, and a bit of a “White Lotus” vibe.

Bref 2**** (Disney +)
No to be confused with Beef 2 :) this is the second season of a comedy series made up of 2-minute shorts about a ±30yo guy and his semi-loser life that was broadcast in France over a decade ago and became iconic. The team reunited to deliver a follow up, when he’s about 40 and still hasn’t come around to growing up, but maybe will? A longer, different format but well done, guys.

Life, Larry and the Pursuit of Unhappiness*** (HBO)
I’m a big Larry David fan. In particular Curb. When the show ended I wondered if he would do something next or relax enjoy his Seinfeld fortune, but he is back with this new “alternate history” show which is quite entertaining.

Rick & Morty Season 9*** (HBO)
Another season is out. The co-creator and main voice actor Justin Roiland was kicked out after Season 6 and I found the show’s style changed for the worse. It feels similar to eating sushi from the supermarket after I lived in Japan for 4 years - the concept is the same but the execution and freshness isn’t.

MOVIES

Not a lot of movies these days but I watched…

Frozen** (Disney+)

• I’m clearly not the target audience and knew very little about Frozen, aside from the title song and the fact that Elsa seems to have become a cultural icon.

• What surprised me most was realizing that the real protagonist is not Elsa, but her younger sister Anna.

• Elsa spends most of the film endangering others while emotionally isolating the one person who refuses to abandon her. I found it hard to believe that Anna would still persist after years of rejection.

• Watching it also gave a new meaning to its title. I had associated “Frozen” with Elsa, but I now think it refers to Anna, who is literally frozen twice by Elsa, as well as to Elsa herself, who is emotionally frozen by fear.

• In the end, Anna is the one who breaks the curse through an act of agency and selfless love. Elsa isn’t the hero. She’s the one being rescued - her title song about self-acceptance hasn’t solved her problem with controlling her abilities.

• In short: Anna is the real hero of Frozen. Too bad she didn’t get a good song.

MUSIC

On occasion, a piece of music impresses me. I hope you enjoy this version of Sultans of Swing as much as I did. Pedro Javier González really knows all the chords ;)

TRAVELS

Florence

• First trip to Florence! Beautiful city, great museums and great food. Some areas crowded with tourists, but that’s in most nice cities these days.

• The Medici were really quite something. Finance and soft power combined.

• It was fun to compare the Bear Cub Thief sculpture from Paris to the goofy Perseus and Dragon (?) in the Boboli Gardens. The latter was apparently finished hastily after a change of sculptors.

• So many Davids! Michelangelo’s (the biggest and most famous) but also Donatello’s and Verrocchio’s which are - in my opinion - much cooler despite being much smaller.

Malta

• A quick visit there for a conference. The country has a magical aura of ancient secrets, blended with a Game of Thrones presence. The language is quite odd, apparently blending Italian and Arabic. Overall worth it.

Amsterdam

• I hadn’t been there in ages and it was fun to see the city again.

• I love maritime museums and theirs did not disappoint. You can’t explain the country without its ships and trade - with all that entails. I also recommend the maritime museums in Lisbon and Hong Kong.

GAMES

Catan

• I played a lot of board games (and pen-and-paper RPGs) in my teens and twenties but I had never played Catan before.

• It is quite an interesting resource-management + negotiation game, no wonder it’s as popular as it is.

• I tend to prefer board games on mobile with AIs because it’s much faster than with humans, though it misses the social and negotiation aspects.

• After installing it I ended up with a mild addiction - even doing some maths to evaluate better the starting positions. I’m cured now.

4. THOUGHTS & EXPERIENCES

On being kind

After listening to a podcast on kindness that left me unsatisfied, I gave it some serious thought, and once again ended up finding that “there are plenty of semantic prisons which do not permit us to think straight” (a wise phrase I found many moons ago in The Human Situation by Aldous Huxley).

Namely, “kindness” is often avoidance in nicer clothes.

• True caring and honesty are not avoidant.

• When kindness consistently avoids courage, perhaps we should call it what it is: cowardice.

Conflict is often necessary

• “Conflict” (another semantic prison) is often conflated with violence, while it has a wide spectrum, from misunderstanding to disagreement to (as a possible outcome) violence.

• From BJJ practice, my view is that even much violence results from a lack of skill and control. A competent BJJ practitioner is controlled and can subdue an opponent without “violence” inflicting damage.

True kindness requires courage, and courage requires practice

All this circles back to the idea that everything is a learnable skill - the classic “growth mindset” - and that the avoidance comes from a lack of practice.

So what is the solution? Practice!

• If you want to learn surfing you don’t just show up in Nazare and try a big wave.

• You start small, and ideally with guidance.

• Standing, timing, then maybe turning, then slightly bigger waves.

So the solution to competence in conflict resolution is to practice on small things with low stakes when appropriate (someone cutting in line? your food tastes bad?), and get reps, so that there is confidence and muscle memory to rely on when the stakes are higher.

On misfortune

• Recently, my friend B. was “fired” by the founders of a startup in which he invested and he had helped full time get to the next level for many months. He was very upset. After hearing his story, I told him “this is likely the best thing that could happen to you” - he had other projects with 100% ownership that energized him, while the startup work was draining him.

• Last week I met a founder who used to be an Olympic athlete. He finished 4th at the Beijing Olympics and his sponsoring dried up. He was devastated. With bleak athletic prospects, he answered a random email from a researcher looking into performance. They ended up growing a genetics testing company together and selling it for 8 figures. Then he got a call that an athlete had tested positive for PED and was disqualified after years of debate, and that he was getting a bronze medal for his past Olympics.

• You might know some version of the old story “we’ll see”.

At close to 50 years of age, my observations have been that it’s impossible to tell if something is good or bad news when they happen.

Even the good news (e.g. getting that coveted job at a fashion magazine) could turn out bad. Staying equanimous, or rejoicing (while knowing it might be all an illusion) is likely the best attitude. And again, a skill to practice.

Let’s whistle while we work.

Until next time,
— Ben