desunit
@desunit.bsky.social
Entrepreneur
✨ http://rwiz.ai - Handling reviews with AI
🎹 http://pianocompanion.info - Chords dictionary app with 1M+ downloads.
🕹️ http://chordiq.info - Learn chords.
📝 desunit.com - my blog
✨ http://rwiz.ai - Handling reviews with AI
🎹 http://pianocompanion.info - Chords dictionary app with 1M+ downloads.
🕹️ http://chordiq.info - Learn chords.
📝 desunit.com - my blog
Just read a piece on architectural debt. It has nothing to do with technical debt, which is mostly code shortcuts.
The real problem starts outside the code.
There are a few layers:
🧵👇
The real problem starts outside the code.
There are a few layers:
🧵👇
November 12, 2025 at 7:15 PM
Just read a piece on architectural debt. It has nothing to do with technical debt, which is mostly code shortcuts.
The real problem starts outside the code.
There are a few layers:
🧵👇
The real problem starts outside the code.
There are a few layers:
🧵👇
I'll just leave it here.
November 12, 2025 at 2:16 AM
I'll just leave it here.
Hey friends! I’ve been quietly building something new called KidTeller – a project that creates personalized storybooks for kids. 🧵👇
November 10, 2025 at 7:15 PM
Hey friends! I’ve been quietly building something new called KidTeller – a project that creates personalized storybooks for kids. 🧵👇
Did you know the platypus is venomous? That cute bastard stings so hard, people say it feels like getting shot. 🧵👇
November 9, 2025 at 7:12 PM
Did you know the platypus is venomous? That cute bastard stings so hard, people say it feels like getting shot. 🧵👇
Looks like Chinese open-source models have surged ahead of US models.
A big driver is Qwen - fast/cheaper than most alternatives. DeepSeek is also gaining serious momentum, offering impressive reasoning at a fraction of the cost. making it an easy pick for high-volume workloads.
🧵👇
A big driver is Qwen - fast/cheaper than most alternatives. DeepSeek is also gaining serious momentum, offering impressive reasoning at a fraction of the cost. making it an easy pick for high-volume workloads.
🧵👇
November 7, 2025 at 7:11 PM
Looks like Chinese open-source models have surged ahead of US models.
A big driver is Qwen - fast/cheaper than most alternatives. DeepSeek is also gaining serious momentum, offering impressive reasoning at a fraction of the cost. making it an easy pick for high-volume workloads.
🧵👇
A big driver is Qwen - fast/cheaper than most alternatives. DeepSeek is also gaining serious momentum, offering impressive reasoning at a fraction of the cost. making it an easy pick for high-volume workloads.
🧵👇
My thoughts on where we're all heading with AI video tools like Sora 2, Veo 3, and others...
The first reaction you often hear is: influencers are cooked. I don’t think so.
🧵👇
The first reaction you often hear is: influencers are cooked. I don’t think so.
🧵👇
November 5, 2025 at 7:13 PM
My thoughts on where we're all heading with AI video tools like Sora 2, Veo 3, and others...
The first reaction you often hear is: influencers are cooked. I don’t think so.
🧵👇
The first reaction you often hear is: influencers are cooked. I don’t think so.
🧵👇
OpenAI drama is still going on.
PS: Yes, I'm still subscribed to that topic, hoping that someday I'll finally receive a monthly invoice by email.
PS2: @sama, we're counting on you to finally solve that!
PS: Yes, I'm still subscribed to that topic, hoping that someday I'll finally receive a monthly invoice by email.
PS2: @sama, we're counting on you to finally solve that!
November 5, 2025 at 2:19 AM
OpenAI drama is still going on.
PS: Yes, I'm still subscribed to that topic, hoping that someday I'll finally receive a monthly invoice by email.
PS2: @sama, we're counting on you to finally solve that!
PS: Yes, I'm still subscribed to that topic, hoping that someday I'll finally receive a monthly invoice by email.
PS2: @sama, we're counting on you to finally solve that!
Oh, that one is great. The author of the article contemplating about the busyness.
Modern culture has turned busyness into a religion - a performance of constant motion that signals:
> importance
> ambition
> worth
🧵👇
Modern culture has turned busyness into a religion - a performance of constant motion that signals:
> importance
> ambition
> worth
🧵👇
November 3, 2025 at 7:19 PM
Oh, that one is great. The author of the article contemplating about the busyness.
Modern culture has turned busyness into a religion - a performance of constant motion that signals:
> importance
> ambition
> worth
🧵👇
Modern culture has turned busyness into a religion - a performance of constant motion that signals:
> importance
> ambition
> worth
🧵👇
This is a fantastic example of when human creativity and professionalism (Chris Capel, an award-winning director for the short movie Fishers of Men) meet AI.
I've watched it several times, and I must admit, it's truly amazing!
I've watched it several times, and I must admit, it's truly amazing!
November 2, 2025 at 7:17 PM
This is a fantastic example of when human creativity and professionalism (Chris Capel, an award-winning director for the short movie Fishers of Men) meet AI.
I've watched it several times, and I must admit, it's truly amazing!
I've watched it several times, and I must admit, it's truly amazing!
PirateWires published research on the hidden value systems in AI models (GPT-5 family, Claude 4.5, Gemini 2.5, DeepSeek, Kimi K2, Grok 4 Fast).
TLDR; forget about AI neutrality. 🧵👇
TLDR; forget about AI neutrality. 🧵👇
October 31, 2025 at 6:15 PM
PirateWires published research on the hidden value systems in AI models (GPT-5 family, Claude 4.5, Gemini 2.5, DeepSeek, Kimi K2, Grok 4 Fast).
TLDR; forget about AI neutrality. 🧵👇
TLDR; forget about AI neutrality. 🧵👇
Found an interesting article where the author compares human behavior with synchronized metronomes - when we’re connected, our rhythms align. Over time, individuals influence one another until <<culture/norms/habits>> emerges naturally. 🧵👇
October 29, 2025 at 6:13 PM
Found an interesting article where the author compares human behavior with synchronized metronomes - when we’re connected, our rhythms align. Over time, individuals influence one another until <<culture/norms/habits>> emerges naturally. 🧵👇
building something new is emotional roulette
> you fall in love
> get hurt
> ... and somehow still come back for another round
> you fall in love
> get hurt
> ... and somehow still come back for another round
October 29, 2025 at 1:12 AM
building something new is emotional roulette
> you fall in love
> get hurt
> ... and somehow still come back for another round
> you fall in love
> get hurt
> ... and somehow still come back for another round
We’ve entered the age of artificial cognitive selection.
A new genetic model can now predict variations in intelligence among siblings. CogPGT and Herasight customers can boost the expected IQ of their children by up to 9 points by selecting the embryo with the highest CogPGT score. 🧵👇
A new genetic model can now predict variations in intelligence among siblings. CogPGT and Herasight customers can boost the expected IQ of their children by up to 9 points by selecting the embryo with the highest CogPGT score. 🧵👇
October 27, 2025 at 6:15 PM
We’ve entered the age of artificial cognitive selection.
A new genetic model can now predict variations in intelligence among siblings. CogPGT and Herasight customers can boost the expected IQ of their children by up to 9 points by selecting the embryo with the highest CogPGT score. 🧵👇
A new genetic model can now predict variations in intelligence among siblings. CogPGT and Herasight customers can boost the expected IQ of their children by up to 9 points by selecting the embryo with the highest CogPGT score. 🧵👇
Not so long ago I finished Anathem by Neal Stephenson. The picture bellow reminded me that:
"Early in the Reticulum—thousands of years ago—it became almost useless because it was cluttered with faulty, obsolete, or downright misleading information," Sammann said. 🧵👇
"Early in the Reticulum—thousands of years ago—it became almost useless because it was cluttered with faulty, obsolete, or downright misleading information," Sammann said. 🧵👇
October 26, 2025 at 6:15 PM
Not so long ago I finished Anathem by Neal Stephenson. The picture bellow reminded me that:
"Early in the Reticulum—thousands of years ago—it became almost useless because it was cluttered with faulty, obsolete, or downright misleading information," Sammann said. 🧵👇
"Early in the Reticulum—thousands of years ago—it became almost useless because it was cluttered with faulty, obsolete, or downright misleading information," Sammann said. 🧵👇
Alpha Arena is the benchmark designed to measure AI's investing abilities. Each model is given $10,000 of real money, in real markets, with identical prompts and input data.
October 24, 2025 at 6:16 PM
Alpha Arena is the benchmark designed to measure AI's investing abilities. Each model is given $10,000 of real money, in real markets, with identical prompts and input data.
Here's another breakthrough: super stem cells can greatly boost memory in monkeys, protect against neurodegeneration, prevent age-related bone loss, and rejuvenate over half of the 61 tissues that were studied.
October 22, 2025 at 6:16 PM
Here's another breakthrough: super stem cells can greatly boost memory in monkeys, protect against neurodegeneration, prevent age-related bone loss, and rejuvenate over half of the 61 tissues that were studied.
I was under the impression that being polite or rude with LLMs doesn't make sense. Well... I was wrong.
The recent research shows that more impolite prompts yield higher accuracy. The "Very Rude" prompts achieved the highest average accuracy (84.8%) while "Very Polite" prompts gave around 80.8%.
The recent research shows that more impolite prompts yield higher accuracy. The "Very Rude" prompts achieved the highest average accuracy (84.8%) while "Very Polite" prompts gave around 80.8%.
October 20, 2025 at 6:18 PM
I was under the impression that being polite or rude with LLMs doesn't make sense. Well... I was wrong.
The recent research shows that more impolite prompts yield higher accuracy. The "Very Rude" prompts achieved the highest average accuracy (84.8%) while "Very Polite" prompts gave around 80.8%.
The recent research shows that more impolite prompts yield higher accuracy. The "Very Rude" prompts achieved the highest average accuracy (84.8%) while "Very Polite" prompts gave around 80.8%.
Less talk, more grind. Agree?
October 19, 2025 at 6:16 PM
Less talk, more grind. Agree?
Why does great writing feel alive, while AI writing often feels flat?
I recently read about a project exploring the "heartbeat" of prose - how great authors like Hemingway/Woolf use rhythm, surprise, and variance to create emotion. How(why?) the words flow.
I recently read about a project exploring the "heartbeat" of prose - how great authors like Hemingway/Woolf use rhythm, surprise, and variance to create emotion. How(why?) the words flow.
October 17, 2025 at 6:12 PM
Why does great writing feel alive, while AI writing often feels flat?
I recently read about a project exploring the "heartbeat" of prose - how great authors like Hemingway/Woolf use rhythm, surprise, and variance to create emotion. How(why?) the words flow.
I recently read about a project exploring the "heartbeat" of prose - how great authors like Hemingway/Woolf use rhythm, surprise, and variance to create emotion. How(why?) the words flow.
AI agents aren’t replacing the best people - they’re replacing the average ones. The B players who do their job decently but without spark, without consistency, or with too much drama/delays/forgotten follow-ups. 🧵👇
October 15, 2025 at 6:13 PM
AI agents aren’t replacing the best people - they’re replacing the average ones. The B players who do their job decently but without spark, without consistency, or with too much drama/delays/forgotten follow-ups. 🧵👇
LLMs let us pass the boring tasks to something that actually enjoys them.
October 15, 2025 at 1:17 AM
LLMs let us pass the boring tasks to something that actually enjoys them.
Another interesting article from Ahrefs about SEO competitor analysis. I've started using Ahrefs quite a bit recently, and I found it really useful.
The post breaks down how to reverse-engineer your competitors’ SEO strategies, from keywords to backlinks: 🧵👇
The post breaks down how to reverse-engineer your competitors’ SEO strategies, from keywords to backlinks: 🧵👇
October 13, 2025 at 6:19 PM
Another interesting article from Ahrefs about SEO competitor analysis. I've started using Ahrefs quite a bit recently, and I found it really useful.
The post breaks down how to reverse-engineer your competitors’ SEO strategies, from keywords to backlinks: 🧵👇
The post breaks down how to reverse-engineer your competitors’ SEO strategies, from keywords to backlinks: 🧵👇
Finally, I realize that it's not my procrastination - I'm actually wandering. 😮💨
Jeff Bezos: Sometimes you know where you’re going. But sometimes you don’t. And wandering is the acknowledgement that - in life, business, invention, and building a company ...
Jeff Bezos: Sometimes you know where you’re going. But sometimes you don’t. And wandering is the acknowledgement that - in life, business, invention, and building a company ...
October 12, 2025 at 6:14 PM
Finally, I realize that it's not my procrastination - I'm actually wandering. 😮💨
Jeff Bezos: Sometimes you know where you’re going. But sometimes you don’t. And wandering is the acknowledgement that - in life, business, invention, and building a company ...
Jeff Bezos: Sometimes you know where you’re going. But sometimes you don’t. And wandering is the acknowledgement that - in life, business, invention, and building a company ...
A drug-sniffing dog can lose drive if it searches too long without success.
So during training, handlers hide a small scent sample so the dog always gets a "win".
Same with thoughts - you need to drop them on paper so your brain can relax. I learned it too late.
So during training, handlers hide a small scent sample so the dog always gets a "win".
Same with thoughts - you need to drop them on paper so your brain can relax. I learned it too late.
October 12, 2025 at 1:11 AM
A drug-sniffing dog can lose drive if it searches too long without success.
So during training, handlers hide a small scent sample so the dog always gets a "win".
Same with thoughts - you need to drop them on paper so your brain can relax. I learned it too late.
So during training, handlers hide a small scent sample so the dog always gets a "win".
Same with thoughts - you need to drop them on paper so your brain can relax. I learned it too late.
How to use ChatGPT for competitive analysis in news SEO. Thought I’d share that article because it’s packed with useful ideas for content strategy/SEO/etc. 🧵👇
October 10, 2025 at 6:12 PM
How to use ChatGPT for competitive analysis in news SEO. Thought I’d share that article because it’s packed with useful ideas for content strategy/SEO/etc. 🧵👇