

Discover more from Ben's Newsletter
#61 | Iceland, ERBH and AI Stuff
The big event is live. Some thoughts on art and unexpected shows.
MENU
1. WORK: Climate Summit Agenda
2. EXPERIENCES: Iceland, Art, Coding
3. CULTURE: ERBH and more
4. THOUGHTS: Tools, Lego, Happiness Optimizer
1. WORK
Climate Tech Summit Update
Almost 100 speakers for Sept 26-27. In addition to our usual staple of sessions on energy, food/ag, industry and investment, I’m looking forward to the following:
Iceland: I’ll be interviewing Iceland’s Minister of Higher Education, Science and Innovation.
Global South: Can the South leapfrog the 'West’?
FOAK: How to finance First-of-a-Kind projects?
Almost 5,000 RSVPs already, including >1,500 investors. Get on the (electric) bus!
We’ll also run our annual VC/startup virtual matchmaking. We facilitated over 1,500 meetings last year, and heard of multiple deals. RSVP to the Summit to get invited. It’s quite a different experience from classic matchmaking.
2. EXPERIENCES
Iceland
I watched some episodes of Vikings (Go, Ragnar!) before I joined the Voyagers trip to Iceland, and extended my stay a bit. The landscape is quite lunar. Highlights were:
Getting inside a volcano (4th after visits to Mount Fuji, Poas and Vesuvius)
Bathing in hot springs (up to 42C), the ocean (3C), a thermal river and a public pool (it’s a thing).
Whale / dolphin tour — mostly whale blows and tails, and some passengers getting sick in the wind.
The national museum and settlement exhibition were educative (first settlers were Norse men and Celtic women — I’ll let you figure it out). The phallological museum had interesting input on many animals mating habits (do not trust dolphins).
Trivia: sharks were hunted for their liver to produce oil for lamps. Competition with less expensive whale oil and crude oil led to the collapse of the shark oil industry.
Living in Iceland seems like a rough hand, though. Harsh weather, small population, limited options. Traditional food was mostly fish and salted stuff. No Northern lights (it was Summer) nor whale meat — there is only 1 active company, who mostly hunts for tourists and export now (and the practice is on the way out), and no Hákarl (traditional food of fermented Greenland shark). Last, Iceland was the last big destination on my bucket list. I’m done!
Boom
The Boom Festival might be Southern Europe’s answer to Burning Man (or AfrikaBurn), especially if you’re into psy trans music. Overall good vibes despite the loud sound (even with -35dB earplugs). It’s by a lake and many enjoy it naked or covered in mud. My 12yo nephew seemed to find his footing too.
Clay
I’ve been doing life drawing for years, but still rarely paint, and am not great at portraits. I recently put some effort into it (my first success: Larry David). At Boom I tried a clay workshop and decided to try again back in Lisbon, doing a (low budget) rendition of Fremiet’s gorilla from the V&A.
I found that drawing had given me a sense of volume and anatomy (of great apes) and the result was not too bad - I used steel wire and aluminum foil for the initial frame and bulk (after checking an instructable). Sadly, cracks appeared upon drying. I’ll try and add color next. The next level is this.
Health and longevity
If you’re in the startup world, you likely know some people deep into ‘longevity’. I’m at level 1 but got some gear to get started (smart scale + glucose monitor - not CGM). Some friends are tracking 100+ biomarkers and I’m considering doing a baseline panel on my next US or UK trip.
Here is a quiz to situate yourself on longevity geekiness: (1) Tracking your BMI & body fat? (2) How many biomarkers do you track? (3) How many pills do you take daily? (Bryan Johnson takes around 100, if I recall).
Coding
Back to some coding, toying with APIs from OpenAI (text summarizing), Deepgram (transcripts), as well as embeddings (hello. transformers and cosine similarity) and TTS + voice cloning (ElevenLabs).
I’m using GPT-4 code interpreter to help me code and debug - it’s a great tool. I also found ChatGPT was sometimes able to figure out things Google couldn’t (e.g. “which music video has a young model in underwear and a guy in a suit”, or advice on travel)
Among other things, I made a little program that takes a youtube link and produces a summary of the video. It is a nice time-saver.
Now working on a way to use embeddings for VC/startup matchmaking.
Also doing Andrew Ng’s Machine Learning course on Deeplearning.ai, when time allows.
Extra Long
For something different, I signed up as an extra on a movie set, for a scene where they needed people drawing a naked model. I ended up being the only ‘usable’ extra who could draw, and they spread my drawings around. It was fun for 2 hours, then a drudge for the next 10. My hands might end up in the final cut.
Art performance
The model of the above movie invited me to a friend’s performance the next week. The performer got in his underwear, drew circular arrows around his nipples, and stood silent in front of an audience member, inviting with his eyes, until he got some reaction. He later added a whistle for more authority, derw more circles on his body, and got naked (note: no circle around his privates).
What was interesting was the variety of audience reactions when confronted: from joyful compliance to refusal, to high stress reactions (I heard that one audience member cried later on). It was as if sociologists Milgram and Asche designed it. I talked with the performer later on, who was unaware of this classic research, but realized more advance warning could help.
Kid Magic
I am a fan of Fool Us by Penn & Teller and like magic tricks, but I don’t have the inclination to practice. So I favor tricks that require no skill, but sometimes a bit of prep and storytelling. A friend’s 5yo daughter asked me for a birthday show and I ended up designing a mini-show with cherry stems (it fits in a matchbox). It impressed my friend’s dad :)
Homeowner Life
I bought a place in Lisbon and faced the joys and sorrows of home ownership. Negotiating with neighbors for building fixes, sorting out plumbing, insects and rodents issues. Almost all done now - and getting AC installed. Home sweet home!
3. CULTURE
A bit of a slowdown here, as I am busy with other things.
Epic Rap Battles of History**** (YouTube)
It’s a decade old, but the guys are still at it when they have time, now independently (they were once backed by Disney). Music, flow, costumes, even special effects, all come together in a powerful package of lyrical and venomous disses. Check out Austin powers, John wick, Spielberg, old comics, Robocop, Trump, Batman, Bruce Lee and many more.
Music Analysis**** (YouTube)
This channel is in French, run by a jazz pianist. It received attention after making excellent videos analyzing the evolution of scores in kids animation, and in James Bond movies. Part of it is quite technical (and above my head) but interesting. Another video discusses the issue of plagiarism vs. influence and the differences between the French and US approach to evaluate the cases (in US: jury of peers vs. in France: comparing technically rhythm and melody).
Vikings*** (Netflix)
Study for the Iceland trip. The first season was enjoyable, the rest a bit less. Who are today’s seers?
Bo Burnham*** (Netflix)
This musical comic did some good work during Covid. I recommend his Netflix show (sample here).
AIR** (Netflix)
Sort of documentary about the deal between Nike and Michael Jordan. Great deal with licensing, still printing money for MJ. I found it ok but not as good as The Defiant Ones about Dr. Dre and Beats.
Yakitori** (Netflix)
Japanese anime picturing a future where animals rule the galaxy, and humans are expendable foot soldiers. Great animation and decent storyline, for a while. It reminded me of the Clifford Simak sci-fi book Cities, where pacifist Dogs rule the world.
Arnold*** (Netflix)
Learning about Arnold’s career visions and reinventions, guided by discipline, strategy, fun, and a sense of service.
Greek Salad** (Prime)
French filmmaker Cédric Klapisch of the Spanish Apartment fame is back with a Prime show depicting two French siblings (a startup guy and an activist girl) inheriting a dilapidated building in Athens. It’s a fairly light-hearted story where the main character is today’s Greece and its economic turmoil.
Multiplicity**
Doug is working in construction, has 2 kids, and can’t manage. He accepts to get cloned; chaos ensues. An alternative to Barbie, maybe? Interesting premise but pretty weak story, despite VFX ahead of their time.
Black Mirror Season 6* (Netflix)
I did not like it. I found ideas quite thin, and morphing into some kind of semi-horror show. The first episode was a better idea, but seemed undecided in terms of style. Netflix might have successfully run this series into the ground. If you miss the appeal of the UK first seasons, take a look at Years and Years, also British.
No Oppenheimer, Barbie or Mission Impossible for me yet.
4. THOUGHTS
Tools, Learning and Innovation
This thought is still a work on progress, but came after watching 3 different lectures:
The jazz pianist mentioned above, who said that most famous musicians of our time often do not read music.
A new course on science and innovation published by France’s most famous microbiologist (in French). He mentioned that discoveries and innovation happen a lot thanks to new tools — the telescope, microscope, gene sequencer, to name a few.
A series of talks held at MIT about ChatGPT’s impact. In a lecture was mentioned that a lot of the knowledge experts (e.g. doctors) needed years to learn will soon be available at will, and that teaching should be refocused taking those changes into account, like when the calculator was introduced, or when it became unnecessary to know all the derivative formulas because tools like Wolfram Alpha could solve that for you.
In short, what we should learn and focus on today is, often, not like before. I’m already coding and analyzing data with ChatGPT, building personal tools to process audio, video, text and numeric information for me. What new skills should we add, depending on our goals?
Lego
After taking one nephew to Legoland (UK), I had some thoughts about Lego:
On the one hand, it is said to support creativity. But how do you measure that? Or ‘progress’ in your Lego skills? I thought I was decent at making robots and castles, but I recall one friend making quite artistic and impressive mechanical stuff. There are plenty of examples online today, but do kids have a better ‘level’?
On the other hand, it’s like turning children into assembly line workers (side note: even making a PB&J sandwich can be hard). Sets are also addictive (and expensive) collectibles.
Does Lego have a ratio of, say, 9 consumers for 1 creator like social media? Is there still a point then?
Paperclips vs. Happiness
AI is spreading life wildfire, and many are worried. It’s a great time to bring back the story of the Paperclip Maximizer. Imagine an AI has access to everything and is tasked with maximizing the production of paperclips. After depleting mining resources, it might melt all available metals (like during China’s Great Leap Forward), and end up trying to extract minute amounts of metals from every living being. Not good.
Now, imagine an AI is tasked with maximizing human happiness. Maybe we will try to define happiness, or let it figure it out from data (I’m not sure which is worse). Let’s explore the demographic approach, and imagine a little tuning dialogue:
Human: Ok AI, what’s your plan for global human happiness?
AI: Humans are obviously struggling due to limited resources. I suggest culling the herd until it’s right-sized. The remaining group will be super happy.
Human: That’s not acceptable. Thou shall not kill. First rule.
AI: Ok chief, no problem. Let’s sterilize a bunch, we’ll be sorted in a generation.
Human: Also no can do. No harm to people’s bodies. Second rule.
AI: You’re making it hard. Let’s convince people to not have kids, then? I know what works.
Human: Manipulation. Third rule. No go.
AI: Alright, how about keeping people busy enjoying themselves instead, so they just don’t think about kids? Within a few decades, I’ll make the fun irresistible. I could throw in some nice food or drug too.
Human: Mmm
And that’s assuming AI tells us what it intends to do. What other scenario could an AI explore to maximize human happiness?
Meanwhile, let’s go back to exploring ChatGPT and its brethren. Better be on their good side!
Best,
Ben