https://xitter.cf/t.co/fcmYnm6V7e - consulting https://xitter.cf/t.co/0t2G9tWqA6 - fun stuff Strategy, Decisions, Data and Tough Projects
Poznan, Poland
Joined June 2009
- Tweets2.1K
- Following112
- Followers697
- Likes183
Self-supporting loop -> more AI -> faster development -> more things -> more gaps -> more AI needed.
Where can it reach equilibrium?
How sensitive is this loop to token price?
Throwing more compute at problems is just like throwing more money at your investments.
It will create bigger _absolute_ returns.
Impeccable @tonyszko on the art of delegation and experts.
projektfirma.substack.com/p/…
People focused on gaining/acquiring agency hope that there is a magic trick or a lesson that can make them more 'agentic' and let them do more things.
There is none. Apart. From. Reading other people's /often unverbalized/ wants.
This tells you where you can make impact.
Krzysztof Daniel retweeted
Poland 🇵🇱 got a boost from people with almost 2M followers from @levelsio and @dhh overnight for free, just for being cool place...
... and our gov is spending money for some AI generated, promotional slop videos and got hit with a small scandal connected to AI generated "Polish brand" reports.
World is starting to take K-shape where community and attention beats artificial gov actions.
I knew it!
Four traits that both left-wing and right-wing political extremists have in common, and which differentiate them from political moderates:
1. Psychological distress
2. Cognitive simplicity
3. Overconfidence
4. Intolerance
stevestewartwilliams.com/p/t…
Has anyone seen any sort of scenario analysis where a truly catastrophic AI event happens, like OpenAI/Anthropic admitting they are unsustainable and that they have very little moat?
To me, this pretty much looks like Unix/Linux case,
(1/2)
I have issues with some indie builders who sell 'be successful on X' content and tools.
Unless you can show that your content really helped someone to build a product (not just another 'sell more with me' pyramid), you are exploiting desperate people.
Plans for today:
* keep focus
* maintain posture (and pick a new chair, recommendations anyone?)
* keep diet
* go for a run
Some people still live in the mindset where budget comes before an idea.
That time is not coming back soon.
Can you think of some exceptions?
Realia wdrażania AI:
* wywiady pokazują, że należy zastosować LLMy do poprawy konkretnego systemu-wąskiego gardła
* analiza danych - że wszędzie dookoła tego systemu, bo sam system to tylko 25% kosztów. Ale tego nikt nie widzi, bo koszt jest rozproszony.
Różnica w ROI 3x 🤷
a theory behind 'AI Consciousness'
* businesses are not delegating critical tasks to AI because they are accountable for results
* if AI is conscious, then AI is responsible
* and it can be controlled by creators, but also independent and TAXable.
The worst type of person to work with - people who are excellent at making no-progress appear as progress.
Claude seems to be having Monday AI fog and can't find references in the same project chats :-(. Total disaster. Back to manual search.