die Kommunikation (Firma u. Medien) Gestaltung ist das Problem. Es fƶrdert wirtschaftliches fomo u. konkurrenzfƤhigen Gedanken und gesellschaftliche Unsicherheit. Die echte positive, innovative Geschichten und verwendungs Potenzialen werden selten berichte.
Posts by Mario Vellandi
Are tech media influencers validly whining about Anthropic or is it just sad rage bait? I understand the updated TOS, enforcement, and 3rd party tool use pricingā¦I just thought the generous subsidies would eventually end and membership tiers would be limited to their products
When I occasionally look at SWE job opportunities in Germany, I get sad how seemingly 95% involves PHP or Java š
An extension to slow down animations and tech demos slowmo.dev
Reach, positioning, and editorial influence for a daily 4-hour āLarry King liveā interview show where guests get an unconditional PR stage. Aināt much diff from a16z buying Turpentine
I find it hilarious OpenAI bought the ātech bro PR netā TBPN. Reminds me of a16z buying Turpentine podcast net, a few years after shuttering a16z future.com
Itās always about buying reach to influence editorial + ad impressions
Last day of NYC vacation before flying home
One thing that I watch out for now is when apps or commands offer help to install plugins/etc, is WHERE it's actually done. Bc right now, it feels like a 'npm install -g' moment and I'm waiting around with a fly swatter
I considered adding beads, but that seems more appropriate at a project-level if doing long term iterative development
At the User level, I now only have the following:
claude-mem (or mem0 for other agents), context-mode (process output saved outside chat), context7 (multidocs search + retrieval), and superpowers
Asking the agent to analyze each skill, plugin, mcp has been enlightening to understand how much is being plugged into the context window and how often (system vs. each turn).
This week I've been learning where agent skills/plugins are installed, optimizing them, and reverse-engineering an image annotation skill from screenshots + descriptions with nano banana pro (very cool!). Now doing various remotion + agent tuts, optimizing github.com/remotion-dev...
I just heard about "autoresearch" by @karpathy; here's a plausible notebooklm summary gist.github.com/mvellandi/84...
I researched 84 different tactics and examples for brands to engage with people in the following categories:
Market Validation
Official Comms
Media and Publishing
Community and Identity
Learning
Interactive Engagement
Innovation
Operational Transparency
docs.google.com/spreadsheets...
In an age of endless media, Iāve increasingly admired great interviewers whether it be Conan OāBrian, Ira Glass, or broadcast news journalists.
It's great, but a tough read at times! I stopped halfway through waiting for a 2nd edition
Itās a bit cringe when I think of tech YouTubers who mix drama farming, sensationalism, rage baiting, mixed with occasional great insights. From the Economist style guide:
cant.stop.laughing
Bingo
shit, cool, now I do. There's just so much going on, plus these damn mega co's are competing on various weekly announcements it's hard keeping up with the actual meaningful
This is like organizational culture, information architecture, and documentation had a baby: handbook.gitlab.com
Yeah, isomorphic JS apps without rsc mental overhead is better. I dig TanStack youtu.be/wkXvv0iJffg?...
At that scale it totally makes sense to get inference capacity for full-featured open models
Coolify rocks! I started using it on hetzner in October and never looked back
The core of swe balances technical, economic, and ethical forcesāa human-centric task that AI currently can't do. Those who master systems theory & high-level complexity mgmt are poised to thrive, as industry shifts focus from "plumbing" & raw code gen to orchestrating complex, large-scale systems
I hear great things about their a/v segmentation and multilingual translation models
I miss the band 311. their sound was fun and only slightly melancholc, ādrugstore cowgirliā aside
Whoah, that's a coffee 'upside-down' world I happily would enter with biscotti
Learned a little Pulumi infra-as-code today. Neat and high level
Iām at an infra conference and a speaker expressed serious concern about a future inability to hire juniors, as ai-related automation will already reduce mid-level workforce