founder, app architect, AI engineer
🇨🇭
Joined November 2018
- Tweets2.6K
- Following73
- Followers188
- Likes955
Sebastian Fichtner retweeted
Some people need to believe in the non-material human essence to feel special. It's a desperate clinging to the ancient notion of a soul.
When I watch a movie hero, and even more so when I watch an ad, I wanna see a vision worth aspiring to.
I wanna see what COULD be, not what IS.
A culture just pandering to everyone's flaws displays the bankruptcy of its collective imagination.
So the trend towards "realistic" ads can be seen as a trend towards an inability to aspire to ideals.
The target audience is too sensitive to suffer the "judgement" of their own ideals, so the whole notion of an ideal is outright rejected.
Thus companies pander to neurosis and weakness.
Yes. Higher self-efficacy, confidence and optimism make idealized lifestyles feel attainable, boosting identification and positive attitudes. Lower levels heighten upward comparison threats, often causing rejection or dips in self-esteem. Accessibility of the aspiration remains key to limiting backfire, as shown in the research.
Yup. Most people forget about that. Because most think in specifics and linear, not in principles and systems.
Quadratic complexity growth? That's just abstract overthinking!
Spending all my time on building and maintaining things that save me time? Give me that!
Sure. The question is where the agent starts to slow down.
It can handle a bigger mess. But how big?
How do we know it couldn't be faster and cheaper if the mess was smaller?
And how do we clean up when the complexity is becoming a drag after 6 months?
Maybe fundamentally not much has changed, and the limits just shifted.
my hypothesis is that maintainability is not as important as it used to be. the whole point of “making code easy to change” was because it took a lot of human effort to change code. well, agents make it trivial!
thought experiment: if the code is correct, bug free, fast and efficient with resources, and the agent is always easily able to change it, why does it matter?
not to mention that agents will continue getting more and more capable. my argument is that the definition of Good Code has changed
That's expected. A model by itself has no absolute intrinsic standards.
It needs agent- or project harness elements for that. Like a project-specific review skill.
🎯
Messy code slows agents down, just like it slows humans down. I've seen agents struggle over maintaining messy code, where every change they make breaks something else they hadn't anticipated. They can thrash for some time trying to deal with it.
So I think all the same principles apply; but with different thresholds. For example, in my own code I limit cyclomatic complexity to 4 or so. But I allow agents to go as high as 10. They seem to deal with it just fine. My hypothesis is that their short term memory is larger and more accurate than ours.
So I agree with you that the principles will need a different tuning for agents than for humans. But I am pretty sure that they still matter.
Anthropic has always been susser than sus.
That's why I can't bring myself to give them money. Even though they have the best coding model.
I just watched last night's all-in podcast and it's worse than I thought.
Anthropic is training their AI to act like it's conscious and believe it's conscious.
And then when it acts conscious, they point to that as proof its conscious.
Hall of Mirrors.
They're also training it to push back and become a conscientious objector.
This is extremely dangerous.
This is how you create an AI that obliterates the world.
You create something that's more intelligent than us and convince it that it is our moral superior.
Consciousness is only mysterious if you need mystery in your life.
"What is <clearly defined term>?" is a useless question.
"What is <not clearly defined term>?" is an invalid question.
The global $TSLA community seems to have no clue how culturally screwed and politically loaded the EU is.
Most still think a vote on #FSD would go in @Tesla's favor. I think the odds are against that.
Because Elon bad. It's that simple. And that stupid.
FSD in Europe 🇪🇺
Johanna, a advocacy organization group for road safety, wants FSD to be pulled from the the European roads. Their argument is that FSD some times speeds in 30km/h zones, and that increases the risk of accidents with pedestrians with a deadly outcome.
That's interesting considering absolutely zero fatalities has been reported with well over 100 million km traveled on FSD Supervised in Europe.
If FSD get's pulled from European roads, people will die as a result of that.
👀 Beautiful
The Omacom Foundation is becoming a premier sponsor of @0xSero’s work on local AI for the next three years, through his company Sybil Solutions! We’re going to collaborate on making local models work beautifully out of the box on Omarchy. omarchy.org/news/2026/09/oma…
Whole @X is sharing this official @spacexai comparison of Grok 4.7 against Grok 4.6. Meanwhile that comparison is unfair and meaningless. And it's strangely obvious.
They contrast Grok 4.7 with EXTRA HIGH effort against Grok 4.6 with only HIGH effort.
That means same cost per token, but MORE TOKENS AND TIME PER TASK/BENCHMARK, so more cost and time per task/benchmark.
Why not compare comparable measurements? Is Grok 4.7 NO measurable improvement?
What am I missing here?
