Laxman KC

Software Engineer | Product Builder

[Laxman KC @TheLaxmanKC]

@pratik_shr Really excited to contribute more on the SkillTrap project.

[Laxman KC @TheLaxmanKC]

Yeah, absolutely! Using voice is incredibly effective.

[Andrej Karpathy @karpathy]

One pattern I find useful for working with LLMs is a nice long ramble session. Sometimes the LLM needs more bits to understand what you're trying to achieve, but you're too lazy to type them. In these cases I like to lean back, switch to /voice and just ramble for like 10 minutes, total mess, anything goes, full stream of consciousness. Sometimes I declare it up top, something like "switching to speech recognition sorry for any typos...". Sometimes I turn it into a small interview of a few turns. But I find that the LLMs are somehow very good at reconstructing long incoherent rambles and often their echo of your own tangle of thoughts comes out quite a bit cleaner than what you started with. The result is that you improve the mind meld and have to correct things less from that point on.

[Laxman KC @TheLaxmanKC]

@athif_shaffy Yeah, check https://t.co/K625X1oy5n

[Laxman KC @TheLaxmanKC]

The hardest part of short-form content isn’t editing. It’s finding the moments actually worth posting.

I’m building AutoClip to turn long videos into reusable content lines, with AI doing the first pass and humans keeping the final say.

https://t.co/K625X1oy5n

[Laxman KC @TheLaxmanKC]

@chamath

[Laxman KC @TheLaxmanKC]

Grok 4.5 ranking #1 is impressive, but solving 9 out of 46 tasks also shows how early long-horizon AI still is. The real race is just beginning.

[X Freeze @XFreeze]

Grok 4.5 now also ranked #1 on the Long-Horizon Terminal-Bench by binary pass rate, outperforming Claude Fable 5, Claude Opus 4.8 and GPT-5.6-sol

Under the strictest scoring metric - where a task counts only if it is fully solved with a perfect reward and zero errors......Grok 4.5 finished clearly ahead of every other tested model

This matters for real-world coding, automation and difficult engineering work because long-horizon terminal tasks require much more than making partial progress

The model has to maintain context, recover from mistakes and successfully complete an entire workflow across hundreds of steps

Grok 4.5 is showing serious strength on complex agentic tasks over extended periods of time

[Laxman KC @TheLaxmanKC]

Learned:

✅ CoAP – Lightweight IoT protocol for devices like smart thermostats.
✅ MQTT – Pub/Sub model for real-time updates.
✅ Built a UDP server with Node.js for IoT.
Excited for real-world IoT projects! 🌐
#IoT #CoAP #MQTT #NodeJS #WebDev

[Laxman KC @TheLaxmanKC]

🚀 Just Launched: Uizzy - A Real-Time Quiz App!

Excited to share Uizzy! Built with React and Supabase, it helps teachers host quizzes and track progress.

Highlights:

💡 Real-time updates
🏆 Scoring
🌐 Responsive UI
🔗 https://t.co/cPdZNod3vu

#WebDev #React #Supabase #QuizApp