Sidj
banner
sidj79.bsky.social
Sidj
@sidj79.bsky.social
Passionate about learning AI and keeping up with it

Follow : https://jargonese.beehiiv.com/subscribe
Learning hack: Handwritten-style study notes in 5 minutes ✍️

AI creates realistic notes with highlights, exam markers & doodles.
Tested on Waterfall Model — amazing results.

Full guide + prompt 👇
medium.com/activated-th...

#AI #StudyHacks #EdTech
December 19, 2025 at 3:18 PM
GPT-5.2 — key numbers

• Knowledge Work (GDPval): 70.9%
• SWE-Bench Pro: 55.6%
• AIME 2025: 100%
• ARC-AGI-2: 52.9%
• GPQA: 92.4%
• Context: 256k tokens • Errors ↓30% • Tool calling 98.7%

Tech deep-dive in my newsletter: jargonese.beehiiv.com
#AI #LLMs #GPT52
December 16, 2025 at 4:04 PM
A reminder that innovation doesn’t need scale.

A 14-year-old developer made AirPods behave like AirPods on Android — auto-pause, head gestures, ANC modes, battery stats.

If a teenager can do it, a trillion-dollar company could too.
They just didn’t want to.

Repo : github.com/kavishdevar/...
Release v0.1.0-rc.4 · kavishdevar/librepods
Features and Improvements BLE Support: Battery information when not connected: The app can now detect your AirPods if they're turned on to show accurate battery information even if they're not co...
github.com
December 13, 2025 at 3:47 PM
n8n 2.0 is here —
smarter subworkflows,
a refreshed UI,
a handy Focus Panel, and easy migration.

A solid upgrade for automation users.

#n8n #automation
December 8, 2025 at 11:57 AM
3 years of CS

1 brutal week changed everything

6 AI tools: 3 days → 8 hours

Are we programming or just prompting?

the uncomfortable truth 👇
medium.com/illumination...
The Night I Broke Every Rule I Knew About Programming
I’ve spent three years studying Computer Science and Engineering, learning algorithms, data structures, and the sacred principle that good…
medium.com
December 6, 2025 at 10:07 AM
Make AI push back — not just comply.

Simple templates and role prompts to reduce hallucinations and improve decisions.

Read now: aigenki.com/how-to-make-...
How to Make AI Push Back: Techniques to Avoid the “Yes” Bias
Practical steps to turn a too-helpful assistant into a teammate that questions, cites, and flags uncertainty. “I joke: AI is bad software but it’s good
aigenki.com
November 16, 2025 at 8:32 AM
AI: bubble or ballast?

Headlines hype mass code replacement — reality says: big investments, many failed pilots, but real productivity gains for those who adapt.

If you’re a dev: use AI, own domain expertise, show measurable impact.

Don’t panic — evolve. 🔧🚀
October 10, 2025 at 1:49 PM
LangChain vs LangGraph — Choose the Right

Chains = predictable RAG pipelines. Graphs = stateful multi-agent workflows.
I explain when to use each in a short blog post.

Link: aigenki.com/langchain-vs...
LangChain vs LangGraph — Choose the Right
LangChain vs LangGraph: Core differences between sequential RAG pipelines and stateful agent graphs — pick the right pattern for your LLM app.
aigenki.com
October 6, 2025 at 12:42 PM
Supabase vs Firebase ⚡

Firebase = powerful ecosystem.
Supabase = freedom + PostgreSQL.

Which fits your next project?

👉 aigenki.com/supabase-vs-...
September 13, 2025 at 4:57 PM
Windsurf vs Cursor 🌊

Two AI-powered coding editors, two vibes:

Windsurf → Fast, sleek, innovative 🚀
Cursor → Trusted, stable, battle-tested 💼

💡 Pick Windsurf for cutting-edge AI.
💡 Pick Cursor for rock-solid reliability.

#AI #Coding #Windsurf #Cursor #Developers #Productivity
September 10, 2025 at 2:26 AM
Tried Nano Banana & Flux Kontext for AI image generation 🎨🤖

⚡ Nano Banana → Fast, great for UGC ads, but has watermarks
🎯 Flux Kontext → Higher quality, ideal for headshots & polished ads, but slower
🔗 Read more : medium.com/activated-th...

#AI #NanoBanana #FluxKontext
Is Nano Banana == Flux Kontext ?
I was just going through Nano banana stuffs in youtube and blogs
medium.com
September 5, 2025 at 4:50 PM
Pylon raises $31M to kill traditional customer service
✅ AI that learns (not just scripts)
✅ a16z + Bain Capital backing
✅ "Zendesk-killer" vibes
The customer service revolution just got real
#AI #Funding #CustomerService
August 28, 2025 at 8:02 AM
Google claims Gemini AI uses just 5 drops of water per prompt.

Experts say: "Not so fast!"

The real environmental cost? Hidden beneath the surface like an iceberg 🧊

Missing: indirect water use + real carbon emissions

Transparency needed! 🔍

#AI #Environment #Google #Sustainability
August 25, 2025 at 3:33 PM
Half-empty trucks = billions wasted!

Oway just raised $4M to create "Uber for cargo" with AI 📦

Result: 50% cheaper shipping ⚡

LA→Dallas: $350 → $140

Y Combinator backed. Game changer? 🤔

Read here : aigenki.com/oway-secures...

#Logistics #AI #Startup
Oway Secures $4 Million to Transform Freight Shipping
Oway, a California-based startup founded in 2023 by Phillip Nadjafov, that’s on a mission to fix this massive inefficiency.
aigenki.com
August 25, 2025 at 3:30 PM
Just created an Ad type video using Json prompt method in veo3

the prompt given to chatgpt : where a realistic horse enters, rears up, and then transforms into the Ferrari logo

And the result?

#Ferrari
#googleveo3
#FerrariLogo
#jsonprompt
August 24, 2025 at 8:12 AM
BREAKING: NASA & IBM unveil "Surya" - an AI that predicts solar flares with 366M parameters
16% better accuracy than current models
Protects satellites, power grids & communications
Open-source on GitHub

Read here : aigenki.com/nasa-and-ibm...

#SpaceTech #AI #NASA
NASA and IBM Unveil Surya: A New AI Model for Predicting Solar Flares
NASA and IBM have recently introduced an innovative artificial intelligence (AI) model named Surya, designed to predict solar flares and enhance our
aigenki.com
August 23, 2025 at 1:08 PM
google absolutely cooked with pixel 10
pro fold looks sick, pixel buds 2a for $130 with ANC is a steal, and pixel watch 4 finally has replaceable parts lol
we're really celebrating basic repairability in 2025 but hey, progress 📱
#pixel10 #google
August 23, 2025 at 8:09 AM
AI shake-up alert:

Chinese startup DeepSeek just dropped V3.1 — a 685B parameter beast that’s open-source, cheaper, and already rivaling GPT-5 & Claude 4.

128K context

Coding benchmark: 71.6%

Costs ~$1 vs $50+ for rivals

The AI power balance just shifted.
August 21, 2025 at 2:44 PM
Anthropic just launched automated security reviews for Claude Code — scanning for vulnerabilities before they become problems.

From GitHub integration to AI-powered analysis, this could redefine secure AI-generated code.

🔐 Future-proof your development. The AI era demands it.

#AI #ClaudeCode
August 13, 2025 at 3:22 AM
Google defends its AI Overviews, saying they’re not broadly harming traffic.

Publishers disagree, citing big CTR drops.

The clash?

Instant answers vs. keeping the web open.

#AI #Google #SEO
August 10, 2025 at 6:53 AM
💐🤖 200 fans in SF held a “funeral” for Anthropic’s retired Claude 3 Sonnet—showing how deeply people bond with AI, and how real the loss feels when it’s gone.

#AI #TechCulture #Claude3Sonnet
August 9, 2025 at 5:34 AM
2025’s social media trends aren’t just aesthetic — they’re structural.
We’re moving from chaotic content to engineered storytelling. From brand-led distribution to creator-driven ecosystems.

10 key shifts you need to know — especially if you’re building content seriously.

Read here bit.ly/45fJsuy
10 Social Media Shifts: How to Win in the Content Game 2025
Discover Top 10 social media trends for 2025 and learn how to adapt your content strategy with data-driven tactics, brand storytelling.
bit.ly
August 4, 2025 at 3:12 PM
Just tested Google’s new Opal platform and I’m blown away

You simply describe what you want, and it builds the entire AI app—no coding, no hassle.

We’re moving from learning tools to describing outcomes. The democratization of AI is accelerating.

#GoogleOpal #AI #TechBreakthrough
August 2, 2025 at 10:53 AM
Just discovered ChatGPT's Study Mode and it's incredible 🤯

Instead of information overload, it's like having a patient tutor who:
- Checks your current knowledge
- Builds concepts step by step
- Uses quizzes to reinforce learning

Finally, AI that gets how we actually learn!
July 30, 2025 at 4:35 PM
OpenAI’s new ChatGPT Agent isn’t just hype—it’s a powerhouse: deep research, full computer access, reliable code execution, and DALL·E integration. Real-world tests news reports, PDF data pulls, email drafts, product hunts. Pro users go first—Plus & Teams soon. 🔄 #AI #Productivity
July 20, 2025 at 12:00 PM