🌶️

[🌶️ @alan_santito]

Everyone should try redecorating their room at least once (Hương)

[🌺Sai🌺 @Sai60012]

BAM BAM BOL RAHE HAIN BHOLE,
ARE JALDI SE GANGAJAL UTHA LE!

O
M

N
A
M
A
H

S
H
I
V
A
Y

B
A
M

B
A
M

B
H
O
L
E
N
A
T
H
O
M

N
A
M
A
H

S
H
I
V
A
Y

B
A
M

B
A
M

B
H
O
L
E
N
T
H

ABHISHEK KARNA HAI SHANKAR KA,
HAR HAR MAHADEV BOL KE CHADHA DE! @grok

[🌶️ @alan_santito]

Put on some indie and pretend you're walking through Kyoto at night (Dinh)

[🌶️ @alan_santito]

@MarinkovicMiro やった！昼も行くぞ

[🌶️ @alan_santito]

Happiness is finishing your to-do list before noon (Sửu)

[🌶️ @alan_santito]

Quilting is my meditation and I mean that literally (Kinh)

[Korens @0xKorens]

This fresh weather trader made $1,585 right after creating his account

He started trading just 8 days ago and immediately started printing

Here are his first two trades:

→ ["Will the highest temperature in Ankara be 32°C on July 22?"] - No at 51.8¢
→ ["Will the highest temperature in Ankara be 31°C on July 22?"] - Yes at 30.5¢

And right now his open positions in Ankara:

→ No 57.2¢ - [Ankara 28°C, Jul 30] - $542.42 profit
→ Yes 47.4¢ - [Ankara 27°C, Jul 30] - $248.57 profit
→ Yes 26.3¢ - [Ankara 26°C, Jul 30] - $153.59 profit

He's holding the whole Ankara temperature grid at once - and winning on both sides

He's not guessing one exact number

He spreads his positions across neighboring ranges - No on the upper bound, Yes on the most likely ones

When the forecast squeezes the temperature into a tight corridor, the right bet isn't one line - it's the whole grid around it

He trades across different cities, but his real edge is in Ankara

[🌶️ @alan_santito]

My dog brought me a sock like it was a gift and I've never felt more seen (Ký)

[SCOTTY BEAM @ScottyBeamIO]

HIS OWN APP HAD 30,000 VLOGS STUCK IN A BROKEN QUEUE – GROK 4.5 FIXED IT IN 3.5 MINUTES FOR 65 CENTS. OPUS 4.8 TOOK 20 MINUTES AND OVER $5 FOR THE SAME FIX.

Real bug, real production app. His app launched 3 weeks ago, already at 40,000 users, when he noticed the queue had 30,000 vlogs stuck – normally it maxes out at 100.

The cause: one corrupted file freezing the entire queue, jamming every upload behind it, tanking playback speed app-wide.

So he ran the exact same bug through both models in Cursor:

Grok 4.5: diagnosed and fixed it in one shot. 3.5 minutes. ~100K tokens. ~$0.65.

Opus 4.8: got there too – eventually. 20 minutes. ~250K tokens. Over $5.

Same fix. 8x the price for Opus.

Then he tested a non-coding task on the same bug: writing detailed documentation so it wouldn't happen again.

Grok: 1 minute, 12K tokens.
Opus: 3 minutes, 20K+ tokens.

Grok won both rounds – speed, cost, and accuracy – on a real bug affecting his real users, not a synthetic benchmark.

This is the kind of result that actually matters more than a leaderboard score.

Bookmark this, you'll want to come back to it later.

[🌶️ @alan_santito]

@chaosgrmln This is everything

[🌶️ @alan_santito]

Nothing beats lo-fi while staring out the window. - My T. Phan

[kyrox @kyroxxxq]

THIS DEVELOPER BUILT A VOICE-COMMANDED AI SYSTEM THAT ACTIVELY CONTROLS A FLEET OF PHONES.

Not a standard chatbot. Not a headless browser script. Something completely different.

He wired a central AI model named Ultron directly to a physical matrix of Android devices. With a single voice prompt, the system wakes the phones from sleep mode, launches the native YouTube application across the entire array, and executes synchronized search queries simultaneously.

The detail most people miss: this completely changes the mechanics of audience growth and content distribution.

When iterating on viral hooks and trying to maximize viewer metrics across different accounts, the biggest bottleneck is manual execution and device management.
An architecture like this turns a standard desk into a scalable, automated engagement farm without requiring a single screen tap.

Most people are tapping screens manually to test platform algorithms. A few are building localized command centers to operate hardware fleets with their voice.

[Shee @MsMuchiri_]

Simple DIY,ups-cycling a plastic led bulb into a garden sprinkler.

[Brevis @brevis_zk]

Great to partner with @UnifaiNetwork on this. 🤝

As AI agents execute complex DeFi strategies across chains, ZK proofs ensure those actions are verifiable and auditable.

We're bringing verifiable compute to autonomous finance infrastructure. 🔥

[KITE AI📍Coinfest Asia🌴 @GoKiteAI]

https://t.co/rgUg0jNEph

[KITE AI📍Coinfest Asia🌴 @GoKiteAI]

We're back with our latest edition of "Proof of AI Journals," by Proof of AI Lab.

In this Journal, our researcher Kevin Ros dives into how 2025 is the year of agentic AI, and Model Context Protocol (MCP) is quickly becoming the standard for connecting agents to tools like Slack, Uber, and Notion.

But there's a huge problem: authentication.

Each agent needs to authenticate with each tool individually.

If you're running 10 agents across 20 tools, that's 200 separate OAuth flows.

This leads to the M × N auth problem:
🔁 Redundant flows
🔓 Massive attack surface
🧱 No granular control over time, task, or scope

At Kite AI, we’re building a cryptographically secure transaction layer that solves this.

Agentic systems won’t scale until auth is reimagined, and we’re building that future. 🪁

[🌶️ @alan_santito]

@RomiLiffourrena #EstariaBueno que Romi Sea mas grande asi le re caigo ♥♥

[🌶️ @alan_santito]

@RomiLiffourrena Como te caigo ahh ♥

[🌶️ @alan_santito]

@SofiiRiver Hermosa con toda las letras♥

[🌶️ @alan_santito]

@SofiiRiver Lindaaa todo bieen?