Deep unconventional -conversion focus- Marketing Insights. Keeping it real. No cold DMs. Nothing.
Joined January 2011
- Tweets 10,063
- Following 339
- Followers 425
- Likes 5,763
1,093 Photos and videos
Pinned Tweet
3 Feb 2023
In 2019 I helped a startup go from 0 to 400k. Within one month.
Something ChatGPT or whatever AI will never, like really never understand
How Funnel Architecting & Google Search Campaigns found our Ideal Customer
6
1
30
15,366
If you know a thing of 2 of automation you will never belief LLM are conscious. They are an assembly of what all automation did compressed. So today I had to bypass a crawl detection, extremely hard. Eventually, I had to fix a mini Python bot that controls a browser via my laptop
1
39
Super humbly grounded, human creativity over machine. So I had to make a button presser automation (captcha). But the button is nuanced, white grey no sharp edges so worked never. So I thought why not force a purple dot via an extension, the local bot instantly grasp after hours!
1
1
28
Here. Lol so tried everything script within browser, self-made extensions everything comes across as a bot. Unless a Python bot runs the mouse as if you. But used 3LLMs no one can ever arrive at the idea of purple dot for contrastš¤£ impossible. Too human. Handsfreeproof lol
22
Jun 14
On the big 5 personality. They should have a criterion for people who just want to say negative or do negative stuff. Not neuroticism. Not antisocial. These people are literally everywhere, always making negative remarks. Or just this negative attitude, even when neuroticism is 0
1
54
Jun 14
Not antisocial because often the most adaptive pro-social integrated people, still have this. Literally everywhere maybe 1 in 5. Could be your boss, coworker, partner, lecturer, or anywhere. It's more: even when completely self-content, they can not easily be okay for others
1
44
Jun 14
Not jealousy either, because it is not that they want what the others have. Sometimes the others they have this attitude towards, do not even have a single thing they admire or aspire to be or have. It is just an attitude at rest they default to. To be negative towards others
37
Jun 14
Physics should create extremely good model of our observable universe for home computing. All planets, stars, gass: highly specific estimated composition & components. Representative solar, galaxy, cluster, supercluster > scales. In our Zoom. Running math & logic simulations
1
52
Jun 14
Not as a one-time project but continuous open-source. Imagine being a physicist, cosmologist, particle physicist. You can run whatever spacetime/model you want. GR as a base layer, or any model, or even unassumed. Running your equations in regions or larger scales simulated
1
42
Jun 14
Imagine how madly we would advance all across the world people running their own models on it, reporting, striking findings. Collaborate. Seeing if it matches observations. I think that must be the future tho. They already have the data. Engineers can fix it. Someone must fund
28
Jun 14
AI can not have ideas. Will always need a directive engineer. So i'm automating finding suppliers, extreme large database. Solved many bottlenecks from 12 to 3 days. From script to Chrome extension. Now I thought, why not backwards parallel? Extremely simple, it has no mind
1
43
Jun 14
It is like not only can it literally stall to a problem, because it has no sense of meaning or priority or wrongdoing. Like days with no forward. But also, does it never think about optimizing its own work kind, because it is extremely rigid forward compute, I can't explain..
1
32
Jun 14
Simple idea, now running on 2 browser one forward the other backwards. When they start overlapping done. Instead of 3 days. Now might be 1.5 days. It started with I swear 20 days lol. If you left it to Fable who kickstarted it. So people will work through things, find loopholes
30
Jun 13
Lol, soccer was my first loveā½ļøā½ļø never made itš¤£ my dream up to 12 was to become a pro, then I realized I was too light. I gave up on lifeš¤£ and became from the streets. Until reality check gave me PTSD in my early twenties
74
Jun 13
Genuinely think mental mode & warmed-up cognition are a real thing. Sincerely realized when I was warmed up patterns became razor sharp super clear. But when I'm in a mental mode like having fun being with the homies. I'm just chilling. So anyone can raise their mode definitely.
1
78
Jun 13
Fable 5 blockage is good marketing PR. 'a model 'so good that the government had to stop it worlwide' 'OMG, what will come next' lol. This anti-climax anticipating the building effect that comes with it will be wild. The short-livedness makes it legendary. Even if it was just Ok.
85
Jun 13
But the best thing ever even for zero technical work. Create a briefing skill before anything. And a rule that forces it that every breakthrough task makes a brief for itself. Local evolution. Does not feel like sci-fi. Just bottlenecks. Or great work is repeated with ease
1
57
Jun 13
For anyone. Content creator seeking trends, one method might be robustly better. You don't want it to re-learn or start from 0. Skill & brief it. Marketers would build a highly technical high ROI landing page. Brief it. Never start from 0. The briefs are for itself. And update it
43
Jun 13
Most people don't use LLm correctly. Let's say you scrape massive data 1 of the things I do. You don't want an LLM to be handsome. Easier to err slower. You want it to break through, find the underlying API Call/fetch structure. Built an autonomous script with a progress bar.
1
42
Jun 13
Making it robust don't mean make it only work. But fail safe. Like temporarily storage in your folder per batch. And saving in cache when something breaks. When all data stored on your device. You again use LLM. To make back-end & great presentation files. 1 CSV, 1 Json, 1 HTML
1
38
Jun 13
This way you safe extreme amounts of tokens. Not only that. Pure code scripts are the ultimate machines! No errors, by something steering & assesing. Pure speed rigid rule following. Did develop myself so its intuitive but you DO NOT need that all! Only think technically like it
33