Canβt believe this! Getting a like & repost from the person who actually inspired me to start my developer journey means a lot. π
Thank you so much, @piyushgarg_dev sir β you just made my day! π
#Grateful#DeveloperJourney#SystemDesign
I wrote another technical deep dive on DEV Community.
This time, I broke down how I built TokenPulse β a Chrome extension that tracks AI token usage across ChatGPT, Claude, Gemini, DeepSeek & Grok.
Read it here:
https://t.co/8b06JqOZVw
#AI#ChromeExtensions#JavaScript
8/ It also tracks ChatGPT, Gemini and DeepSeek.
Free. Open source. 30 seconds to install.
https://t.co/MlvOVy1H95
If you use Claude daily β install it before your next session gets cut short.
7/ I built a free Chrome extension that shows both limits in real time.
- Exact 5-hour % used
- Exact 7-day % used
- Countdown to each reset
- Works on https://t.co/LKd7s7Yh8g β no API key needed
5/ The dangerous pattern:
You hit the 5-hour limit.
Wait it out.
Come back fresh.
Hit the 7-day limit 20 minutes later.
No warning. No countdown. Just stopped.
3/ The 5-hour limit resets on a rolling basis from your FIRST message.
Sent your first message at 9am?
Resets at 2pm β not midnight, not on the hour.
Most devs think it resets at midnight. It doesn't.
2/ Claude has TWO rate limits most devs don't know about.
- 5-hour session limit
- 7-day weekly limit
They reset independently. You can max both on the same day.
Built TokenPulse: a free Chrome extension for Claude + ChatGPT.
β Live token tracking
β Claude limit timers
β Alerts before limits
β Local-only data
https://t.co/hN2Qz6rRIe
Why is @Medium sliding my article page left to right automatically while I'm reading? Has anyone else hit this? This UX is broken and nobody's talking about it.
LLMs are smart, but forget everything.
New chat = reset. No memory.
Not a bug β context window limit.
So Iβm building:
β store chats
β retrieve context
β feed it back
Goal: memory-aware LLMs.
#AI#LLM#GenAI#BuildInPublic
Learning how phones and laptops share files without a USB was a legit mind-blow.
One tap = discovery, secure handshakes, encryption, data transfer β all in seconds.
Everyday features hide some seriously impressive engineering.
#SoftwareEngineering#SystemDesign#Networking
Learning how LLMs actually work changed my perspective on AI.
They arenβt magicβjust well-engineered systems trained to recognize and predict language patterns, turning messy human input into usable output.
#AI#LLM#ArtificialIntelligence#MachineLearning