THIS ONE GPT-6 ASTRA WORKFLOW MAKES 3D THAT LOOKS HAND DRAWN
you type one clear sentence and it builds a wireframe scribble figure that overlaps your headline
no modeling anything from scratch
you take any GLB, drop the mesh and keep only new THREE.WireframeGeometry(geo) with a thin LineBasicMaterial
then you jitter every vertex a little on load, position.array[i] += (Math.random() - 0.5) * 0.04
that tiny jitter is what turns a clean wireframe into a hand drawn scribble
black lines on white, opacity around .5, so overlapping lines build density where the form is thick
rotate 0.002 per frame on Y only, no orbit controls, the pose does the work
put the headline on top with z-index and let the figure overlap the letters
that overlap is the whole composition
the trick is asking for the behaviour, not the vibe
wireframe scribble figure, black on white, slow Y rotation, overlapping the headline beats make it look cool every time
tomorrow I drop the full build so you can ship this look from an empty file
HATH JODKAR KHADI HOON TERE DWAR, MERI MAA,🚩🙌
P
O
O
R
I
K
A
R
D
E
M
U
R
A
D
E
N
E
K
B
A
A
R, MERI MAA. 🙏
M
A
I
N
T
O
C
H
A
D
A
R
B
I
C
H
A
U
N,
M
A
I
N
T
O JYOT JALAUN,🙏
TERE DAR SE KHAALI NA JAAUN, MERI MAA. 🪷 @grok
Shiv Shankar Bholenath,
Your damru plays upon the mountains.
O
M
N
A
M
A
H
S
H
I
V
A
Y
Om Namah Shivay🪔
O
M
N
A
M
A
H
S
H
I
V
A
Y
OM Namah Shivay🙏
O
M
N
A
M
A
H
S
H
I
V
A
Y
Om Namah Shivay ⚜️
O
M
N
A
M
A
H
S
H
I
V
A
Y
TERI SADA HI JAI MAHADEV @grok
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
Allah
♥️ I LOVE ALLAH♥️
#خاتم_النبیین_محمدﷺ#درود_وقرآن
🙏Om Namah Shivaya 🙏🏵️
O
M
N
A
M
A
H
S
H
I
V
A
Y
A
🏵️
🔱 Har Har Mahadev 🔱
H
A
R
H
A
R
M
A
H
A
D
E
V
🏵️
🔱 OM NAMAH SHIVAYA 🔱
🏵️
O
M
N
A
M
A
H
S
H
I
V
A
Y
A
🏵️
🔱 Har Har Bole Namah Shivaya 🔱
🏵️
H
A
R
H
A
R
B
O
L
E
N
A
M
A
H
S
H
I
V
A
Y
A
🏵️
🙏 @grok Har Har Mahadev 🙏
ZTE and MTN South Africa have taken a major step forward in advancing AI‑native green network innovation. The commercial validation of ZTE AIR RAN EnergyTurbo in Western Cape delivered a significant improvement in site energy savings — from 2.8% to 13.5%. 🔋 Details:
A BUILDER TURNED A 5-BAY ORICO VAULT PACKED WITH MIXED DRIVES INTO AN 110TB LOCAL SECOND BRAIN POWERING HIS ENTIRE AI STACK WITH ZERO DEPENDENCE ON THE CLOUD.
HE OWNS EVERYTHING. NO SUBSCRIPTIONS. NO RENTED INFRASTRUCTURE.
Obsidian acts as the knowledge hub, while n8n automates the entire pipeline ingesting Claude outputs, testing AI agents, and archiving useful patterns directly onto local storage.
The hardware advantage
Five hot-swap bays support a mix of SSDs and HDDs. Each drive locks securely into place, while seven RAID configurations can be selected with simple DIP switches. One power cable keeps the whole setup running quietly.
Sitting beneath a Mac Mini, the vault stores every experiment, dataset, and note inside one searchable Obsidian knowledge graph. n8n workflows run around the clock without API throttling or per-execution fees.
The evolution
Week 1: 110TB online.
Month 1: Automated pipelines running.
Month 3: First agent turned into a product.
Month 5: The system becomes self-sustaining - ingest, test, archive, repeat.
Old drives get a second life. Lockable bays protect the data. RAID adds resilience.
Local-first infrastructure transformed a pile of scattered files into an autonomous AI system.
Because a second brain shouldn't just store knowledge.
It should compound it.
A $440 DELL OPTIPLEX MICRO CAN TURN A HOME RACK INTO OWNED AI CAPACITY FOR REPEATED PRIVATE TASKS
The owner starts with the Dell under a small rack.
The Raspberry Pi is the familiar first home-lab box.
Not this: treating either box as a cloud replacement.
Instead this: give one machine one repeating job.
Private document sorting is a better target than hype.
Local transcription can run without sending files away.
A local assistant can search notes on the same network.
The value is not one huge model answering one prompt.
It is a queue of small jobs that keeps returning.
Not monthly seats for every repetitive internal task.
Instead owned capacity that stays available after setup.
RAM is the first ceiling, not the rack aesthetic.
Choose the Dell when the workload repeats every week.
Choose the Pi when learning Linux matters more than speed.
Bookmark this before your next cloud bill.
Follow for local AI builds that actually change the math.