Daniel Dynesius
banner
danieldynesius.bsky.social
Daniel Dynesius
@danieldynesius.bsky.social
🏞: Founder (data/ai)
🌞: Psy, Phil, gamedev, poetry, Responsible AI
💭: Swedish Covid Strategy is Unethical
👍: honesty+integrity
📈: https://covidradar.vercel.app
🌐: https://www.asciendo.ai
One thing I'll say is, I think it's important that we still write our thoughts and words in our voice, such as it is. Putting our thoughts, feelings and ideas through machinery erases what we are.
October 7, 2026 at 4:33 PM
Technological progress increases capabilities. Without good judgment, obtained through humanities and art, society regresses.
October 7, 2026 at 3:56 PM
Is it obvious to you why AI cannot be allowed to (recursively) self-improve?

Here’s why I think it should not be allowed:
astraatlantis.substack.com/p/misaligned...
Misaligned AI Should Not Be Taught To Self-Improve
Doesn’t seem wise to empower an entity that doesn’t care society
astraatlantis.substack.com
October 7, 2026 at 11:30 AM
Wrote a brief argument about the necessity of #regulation for powerful #technology

astraatlantis.substack.com/p/a-case-for...
A Case for Stopping Frontier AI Research
An inductive argument for international safety regulation
astraatlantis.substack.com
September 29, 2026 at 5:56 PM
GenAI seems on track to become the most profoundly disempowering invention in history.
September 15, 2026 at 10:32 PM
1/2 One way of imagining the change of current society, to society in the advent of #AGI - could be made with an analogue to a society of children deciding and governing, but then suddenly many adults beginning to arrive..
#ai
September 8, 2026 at 5:51 AM
I've noted that, interviewing for various data/programming positions, that alot of people seem to avoid to talk about #AI in a very similar way to #covid . It's like, we know it's a factor but we don't like it and don't want to talk about it.
July 16, 2026 at 12:08 PM
In the age of LLMs, apart from energy & water, there are vast data needs. Essentially, all data from history, today & onward, is needed for the race towards AGI. It's a premptive landgrab that weakens privacy, intellectual property, extracts utility & sovereignty of individuals.
June 30, 2026 at 2:33 PM
One aspect frontier LLMs is the security gap detection and potential to exploit them. If such capabilities are large enough, then digitalised societies, operations and economies find themselves under a large risk of disruption of their necessary function.
June 28, 2026 at 5:52 PM
1/2 "It [Ebola] doesn't spread through the air like COVID and influenza" - Sweden's Deputy State Epidemiologist, on the latest outbreak. Remarked almost in passing, as if it was common knowledge: COVID's transmission route changes from droplet to airborne..
May 20, 2026 at 12:57 PM
When im #vibecoding as an experienced engineer, things get created relatively quickly (after multiple iterations and dialogue flow). But as projects grow, it becomes much less manageable to update and extend than human written code.

You essentially give up control for speed.
April 23, 2026 at 4:05 PM
Har tänkt på folksinnet i #Sverige 🇸🇪 (som ligger på ett kollektivt omedvetet plan). Har övervägt ord som "Exceptionalism", "Statsnationalism" etc. Men jag tror det som fångar folksinnet bäst vore en term som "Helig statsnationalism"

#sociologi #sociology
April 2, 2026 at 5:41 PM
I'm wondering if the reason for demands of ID/age verification on internet/social media etc is because the Ad revenue model is threatened by the advent of agentic AI. Advertiser simply don't want to pay for bot clicks, and ID verification would save the ad businesses e.g. Facebook/Meta.
March 18, 2026 at 1:54 PM
Reading some Substacks I begin to wonder, how much of this is actually AI written or edited? If this continues, it creates a dilution of thinking even in the 'intellectual' spheres of society.
March 18, 2026 at 8:10 AM
When evaluating the #morals of the distant past, like from the medieval or antiquity – we look at them as rigid and dogmatic. That is probably true. Yet we forget that tech & culture didn't change as fast as now – making their morals at the time *seem* to be timeless.
March 16, 2026 at 2:02 PM
I wonder if the main reasons ancient greek philosophers are admired and remembered today is that their thinking was broad enough to encompass more than knowledge in their value systems. Their thinking related to the society, nature & being. Timeless questions. #philosophy
March 6, 2026 at 5:34 PM
I think it makes sense to ask the question, if large corporate investment should be considered political.

I think it is, especially when they transform society.
February 21, 2026 at 3:28 PM
The rolling out of AI work automation is called 'inevitable'. At the same time, that tech takes away people's potential to shape a good future for themselves. It's important to know that there's a difference between technological and human progress and society.
February 14, 2026 at 2:18 PM
Reposted by Daniel Dynesius
Read what NIH workers had to say about their new boss last year:

www.importantcontext.news/p/out-of-his...
“Out Of His Depth,” “Sold His Soul,” “Clueless”: NIH Staffers Speak Out About Director Bhattacharya
Widespread dissatisfaction over the NIH’s “continuous free fall” has people speaking out.
www.importantcontext.news
January 29, 2026 at 10:11 PM
You know, when #society speaks about digital addiction, endless scrolling etc.

Its nothing new. Essentially its just a individual hedonistic pleasure and gratification over meaning and purpose.

To say it more briefly;
The difference between Virtue and Vice.
January 6, 2026 at 1:21 PM
It's been a bit on my mind. The way people orient in #society is driven by how others behave, believe, think and feel.

Its still this simple fundamental condition: social cohesion, and wholeness found in the tribal condition. It's saught.
December 29, 2025 at 12:39 PM
For most, I think, genuine social connection is what is experienced as most meaningful.

Tech weakens these connections. Leaving most in a place with less meaning, warmth & purpose.

Focus on individual success in place of social usefulness, displaces meaning in a similar manner
December 18, 2025 at 3:38 PM
"You will not lose your job to AI, but to a person using AI." This phrase has been used to push AI adoption the last couple of years.

To me, it seems far easier to replace "thought"/work of someone using AI, compared to one who doesn't (all other things being equal).
December 15, 2025 at 6:26 PM
If #AI is trained on Internet data as its basis. How can we just assume it will be morally good & ethical by default?

We simply can't.
December 13, 2025 at 8:50 AM