If you only read one thingđŹ Marketing High Tech Without Talking Tech (5 minute read) Youâre not a marketeer, so how is this interesting? You are a marketeer! You talk to many people outside technology, you market yourself and technology. And when I listen to my CTO clients, they talk too much mumbo-jumbo that no one understands. As the article writes this âalienates non-technical audiences [and] fails to communicate real-world valueâ When talking to the CEO, for a project, problems, more budget, âFocus on Benefits, Not Featuresâ, âSolve a Problemâ, âUse Analogiesâ and âTell a Storyâ (youâre the storyteller in a company, why the moon rises and the sun goes away). I love the practical exercise:
For more success getting what you need and want. https://slimsaas.com/blog/marketing-high-tech-products
Stories Iâve enjoyed this weekđ The Art of Interviewing Your Interviewer to Uncover a Companyâs True Culture (5 minute read) Not enough people see interviews as a dialogue. Your goal is not to get a contract (as the candidate), but to get a great job. Too many people go threw and interview, succeed and then what? They think âWhat have I done, I donât want to work here!â The article has some great questions to ask your future employer, to find your dream job ;-) https://praachi.work/blog/questions-to-ask-in-a-job-interview Googleâs former CEO says the tech giant is losing out to OpenAI and Anthropic because staff are working from home (4 minute read) Not sure if this is the reason that Google lacks in AI (Also think itâs a PR problem). But there are many good points in the interview you and a must-read, even if you disagree with the points. My guess would be, remote does not work for Google because they do it wrong. No one talking to each other outside meetings. You need to add ways that people can talk and interact outside of meetings to create creativity and serendipity and motivation (We can do this! And we want to!). So yes, doing remote wrong vs. office is worse. Duh! The complete (AI, probably not Gemini though, ha!) transcript (with errors) can be found here, worth a read: Github, trust me, link too long đ§Ș The Elon / Trump interview on X starts with an immediate tech disaster (19 minute read) So Elon Musk fired a large majority of Twitter employees, now for the second time a large video stream blows up. The big-layoff experiment keeps being suspenseful, no end in sight. If you had planned to follow Elonâs model to layoff people and cut tech stuff, wait for now Iâd say. But it is an interesting experiment, how many people do you need to run a chat app (When I was at eBay, there were ~30.000 people there running a website, I thought, huh, that are quite a lot of people to run a website). https://www.theverge.com/2024/8/12/24219121/donald-trump-elon-musk-interview-x-twitter-crashes Do Large Language Models Know What They Donât Know? (5 minute read) Do they? âOur extensive analysis, [âŠ] discovering an intrinsic capacity for self-knowledge within these modelsâ. My AI experience around this is hit-and-miss. Today ChatGPT wrote Go code for me to add a border to every SVG in my upcoming book. No problem there, it knew what it knew. Then I had a conversation with ChatGPT about backwards compatibility of some code, and it insisted it was right, when clearly it wasnât. Several prompts later I still couldnât convince it that it was wrong. I didnât use the code. Does it know what it doesnât know? But then again, this also happens when I talk to engineers (and Iâm sure they feel the same about me sometimes). https://arxiv.org/abs/2305.18153 Thereâs An AI (9 minute read) There is an AI for everything, great overview and directory of AIs. Very useful. Postgres Powered by DuckDB: The Modern Data Stack in a Box (26 minute read) Today DuckDB seems to be everywhere. And it is now in Postgres âOne of the most surprising and disruptive aspects of DuckDB is that it is taking fast analytics away from the complex world of distributed systems back into the much simpler single-machine realm.â Single machines are so much easier to run compared to a Spark cluster for data processing - if you donât have large devops and data teams. Once we lost the only Scala developer, ouch! It does look like you now can run Postgres with DuckDB on large data sets in S3 with SQL knowledge alone on one machine. VoilĂ ! https://www.crunchydata.com/blog/postgres-powered-by-duckdb-the-modern-data-stack-in-a-box Go big or go home (32 minute read) Having worked (and founded) in VC funded startups, and having founded (and sold) a bootstrapped one, the real difference is risk appetite. VC funded companies plan for lawyer fees of several ten thousand dollars, which would just kill any bootstrapped company. With VC money, you not only can, but VCs want you to be much more aggressive (lawyers included!) https://www.businessinsider.com/startup-success-secret-vcs-founders-nate-silver-on-the-edge-2024-8 Ex-Google CEO says successful AI startups can steal IP and hire lawyers to âclean up the messâ (22 minute read) Proofs my point about VC money and lawyers. From the same interview, but youâve read the transcript already, didnât you? https://www.theverge.com/2024/8/14/24220658/google-eric-schmidt-stanford-talk-ai-startups-openai Move Slow and Fix Things (10 minute read) The opposite take. No money, move slow and fix things (vs. move fast and break things). The crafter
inside me loves this take, the startupper in me wants VC money. Which often confused âmyâ employees,
I can hold two contrarian opinions at the same time (e.g., CTO vs. Scrum Master, being both at the same time). https://endler.dev/2024/move-slow-and-fix-things/ All GitHub services are experiencing significant disruptions (31 minute read) Next time you have a website-down and incident, and your CEO is angry, tell them even GitHub (some say it has increased since the Microsoft takeover, but Iâd say itâs growth, people, not users) has incidents. Besides that, the small post-mortem is interesting. https://www.githubstatus.com/incidents/kz4khcgdsfdv How We Survived 10k Requests a Second: Switching to Signed Asset URLs in an Emergency (15 minute read) Do you use signed URLs for assets? No!? https://hardcover.app/blog/how-we-survived-10k-requests-second-switching-to-signed-urls The myth of the Product-Market Fit. (11 minute read) I disagree, of course. PMF is mucho important. But too many people do not focus on PMF, so it never comes, but slow acceptance arrives. And they confuse PMF with channel-product fit (where the hockey stick comes from, read the Traction book!). As again does the article. But you can learn from reading it nevertheless. https://blog.nishantsoni.com/p/the-myth-of-the-product-market-fit Rule of the weekSomething to try the upcoming week  Join the CTO newsletter! | |
