This week, I was back mentoring the new batch of teams in the WHU Accelerator again – for the fourth or fifth year in a row!
“The WHU Accelerator is an equity-free intensive program for WHU students, alumni, and staff designed to provide rocket fuel for your new ventures.”
“WHU – Otto Beisheim School of Management is a leading German business school with an exceptional national and international reputation. WHU offers academic programs and continuing education for executives throughout their career.”
I’ve got some news as I can’t sit on the fence any longer in light of the geopolitical developments.
I’ve been searching for ways to get involved, to meet and connect with likeminded people who also wants to do more than to silently observing our European values, democracy, and security eroding, people who want to get involved, people who actively want to do something.
However, I couldn’t find anything – so I’m starting the Defence Tech Meetup – I hope to build a community of people willing and able to contribute to protecting democracy, European values – our future – by helping to build the next “Made in Europe” defence technologies in any shape, way, or form.
This is for doers, hackers, developers, founders, tinkerers, engineers, builders, and people of all sorts to meet and find the people you need to build it with, find other ways to contribute, find real actual problems and needs to build solutions for.
In short, the goal is to gather like-minded people to create a community where you’ll find whatever you need to get involved – and once you’ve gotten involved, that it will serve you as a place for help and inspiration to help keep you going.
I hope it can also serve – in a very modest way – to help make building defence technology in Europe, for Europe and our allies, gain the urgency, attention, recognition and respect it deserves (especially in Germany, where for obvious historical reasons the defence industry has been a controversial and touchy subject).
Get in touch if you want to start a Defence Tech Meetup yourself – wherever you are. I’m happy to share resources and contacts, help you set it up and get it going. After all, this is an idea, it’s not about a specific person or a specific place. The more, the merrier. It’s all open source and free – take it and run with it.
Meetup #1 is already scheduled for May 14th 1830 hrs CET at STARTPLATZ Düsseldorf and the first speakers are already secured:
Speaker 1: Stavros Messinis(Greek, founder “Smart Flying Machines”, exit to Delian Alliance Industries, formerly Lambda Automata)on founding a defence tech startup to exit in less than two years.
Speaker 2: Orrick, Herrington & Sutcliffe LLP (Orrick, one of Europe’s Leading legal firms for innovation & technology) on “How to stay out of jail (when building defence tech in Germany)”.
How can you help? If this is for you, register with the Meetup group and show up for the meetups! Not for you? Help get the word out, share this publicly wherever you think appropriate, feel free to forward to companies, organisations, individuals you know this would resonate with – that would go a long way!
The Meetup Format
Introduction: New members introduce themselves, background, skills, complete with ask (what do you want out of the meetup) and give (what can you contribute to the meetup), 60 seconds standup
Talk 1 (e.g. current challenge(s)/need(s), updates or perspectives, on premise or beamed in over Zoom, defence industry representative, NATO, national defence forces, frontline allies, analysts, authors, organisations on the ground, etc)
Q&A session with speaker(s), open mike, first comes…
Mingle & Network, intermission
Talk 2 (e.g. “lessons learned”, from defence tech entrepreneurs, engineers, industry, defence forces, elected officials, frontline allies, etc)
Q&A session with speaker(s), open mike, first comes…
Vote, show of hands: Are we doing this meetup again?
Suggestions, if yes vote, topics & speaker(s) for the next meetup, improvements, etc collected.
Mingle & Network, open end
All meetups are under Chatham House Rules: Anyone who comes to a meeting is free to use information from the discussion, but is not allowed to reveal who made any particular comment in public – unless explicitly allowed to do so.
No media recording allowed: Unless explicitly permitted by everyone captured on the respective media.
Recently I stumbled upon some oldposts sharing what I had been reading (between 2010 and 2012). It’s been a while since then – and I don’t think I’ve shared much of what I’ve been reading (or re-reading) since.
On the one hand, it seemed to be a prohibitive long list to write by hand – so I haven’t bothered to do so (until now). On the other, since all of my reading purchases have digital receipts of one kind or another, creating an updated list of reading seemed like a tempting opportunity to test how far AI has come (or not), using a real life menial task consisting of several steps, including several sources, and different types of data – something so tedious and time-consuming (and for no payoff to speak of) that it needs outsourcing of one kind or another to actually get done.
TL;DR
Coaxing ChatGPT took an inordinate amount of trial and error to get to do anything useful (and consistent) at all, but after a lot of painfully crappy results I got to a point where the process was somewhat manageable, making the process significantly faster than using manual labor only – but still requiring some manual labor nonetheless.
When ChatGPT was asked to do the work, it refused to. Constantly. And when it did do some work, it lied. Randomly. And when it actually produced usable results, it worked slower than an intern with a hangover. Also, doing random spot checks, I found that books were missing, inexplicably dropped along the way, some summaries were hallucinated, some links where swapped with placeholders, etc.
Key Insights
It helped a lot to break down tasks into the smallest single step, do not bundle steps in a chain (do not use and and and), complete one step, move on to a new instance for the next step, prompt for new task and input output of previous step in this new instance. Counterintuitively, it seems the less ChatGPT 4o knows about the whole, the better the results. So much for artificial intelligence.
Also, try repeating the same menial task with same prompting in multiple different instances – it doesn’t matter how or what you prompt, Chat GPT *will* (sooner or later) go off the reservation regardless. It might just suddenly work in one of the instances (by using exactly the same prompting that failed in 10 of the other instances) – it’s a crap shoot that makes for an entirely sucky user experience.
For identifying the books I’ve read since, I dug up the digital receipts from Audible, Amazon, and Apple and made screenshots of them for speed. However, 4o sucked at OCR. Not so much the fidelity of the results but for speed, or complete lack thereof. ChatGPT took ages to complete a batch of images, broke down or provided hallucinations if the batch was larger than say 5 images – and even then it randomly failed or lied. So I ended up just using manual select, copy, and pasted the relevant texts to ChatGPT via Apple’s vanilla “Photos” app for OCR of the same screenshots because although that included manual work, it sped up the processing more than an order of magnitude – and the results were consistently correct.
To wit: OpenAI is the most overvalued startup in history – I keep asking myself why I pay them; ChatGPT as a product experience is (still) very bad.
The links to Amazon and Audible do include affiliate tags. (If you don’t want to feed the Bezos machine, you probably know how to get it somewhere else). However, I’ll never get back even a fraction of the time it cost me to wrangle ChatGPT to do some actual boring menial work.
Now what have you read that I haven’t (but should)?
Read on below for the +365 items that made it to the list.
Recently I had the honor and pleasure to be invited to hold a keynote speech for Alumniportal by Deutscher Akademischer Austauschdienst (DAAD).
I spoke a bit about five things I wished that I had known before starting out as a startup founder.
Here they are – in no particular order.
ChatGPT’d screenshot from the online event
About the DAAD:
The DAAD is the world’s largest funding organisation for the international exchange of students and researchers.
Since it was founded in 1925, around 3 million scholars in Germany and abroad have received DAAD funding. It is a registered association and its members are German institutions of higher education and student bodies. Its activities go far beyond simply awarding grants and scholarships. The DAAD supports the internationalisation of German universities, promotes German studies and the German language abroad, supports countries in the Global South in building and improving their higher education systems and advises decision makers on education, foreign science and development policy.
Recently I was invited by students at the prestigious (often ranked as #1 for business degree universities in Germany) private university of WHU Otto Beisheim School of Management in Vallendar, Germany to mentor at their 3Day- Startup (sorf of a Startup Weekend clone) event.
And the winner was… Team Care Screenshot
I’ve also been working with WHU in Düsseldorf for some years now, mentoring at their accelerator program, but I had not been to WHU Vallendar since I was invited to Idealab in 2013, so it was nice to be back again.
