§wam
@haveYouMetSam_
Followers
165
Following
11K
Media
98
Statuses
788
Software Engineer | Love the dopamine kick that learning new tech gives me 👾
Joined January 2020
Was missing QuizUp — so I built my own. (This is why I love being an engineer) 👉 https://t.co/JZ06DNOWx4 🚨 QUIZZI v1.0 is live! ⚡ Pick your favorite topic ⚔️ Challenge a friend in Battle Mode 🎮7 rounds per game with 2x points in the final round Let's get trivial! 🧠🏆
2
1
3
happening literally in between a hackathon 🥲
Nearly 125,000 San Francisco households and businesses lost power Saturday due to PG&E blackouts that brought a massive portion of the city to a halt — darkening businesses and homes, causing traffic jams and forcing BART trains to bypass at least two downtown stations.
2
1
4
Gemini flaunting in Tamil at Devhouse SF! @DynamicWebPaige totally was not forced to do this!
1
3
14
This clip has shown up on my timeline like 10 times today, and I let the whole thing play every single time
You made it viral! So now it’s here, Sher-E-Baloch's absolute 'Bang-er'! 🔥 Book your tickets. 🔗 - https://t.co/cXj3M5DFbc
#Dhurandhar Blazing In Cinemas Now! @RanveerOfficial #AkshayeKhanna @duttsanjay @ActorMadhavan @rampalarjun #SaraArjun @bolbedibol @AdityaDharFilms
26
303
4K
Taking Ls with such confidence from even actors and engineers now 🤣
@striver_79 @aifiesta I don’t think you understand our platform We have a Super Fiesta mode also, where in a single chat, our algorithm decides which AI to give your prompt to. We needed to weigh and test Gemini 3 compared to other models to decide how to place it in this mode.
0
0
0
When I'm on my 839472th "Please fix the code" and Ajay Devgn pops up on my screen with a Vimal ad 😭😭😭
🚨 OpenAI is exploring the introduction of advertisements in ChatGPT as a new revenue stream.
0
0
0
life hack: stop trying to be cool. be nerdy and obsessive about the things you love. enthusiasm will get you farther than indifference.
206
30K
163K
Rage baiting at its finest. (The engineer in me quite amazed though 😆)
Chad IDE (@cladlabs) is the brainrot code editor. AI coding takes 1–5 min between prompts—too long to ignore, too short to start something new. Chad integrates your brainrot (X, IG, Stake, Tinder, etc) into your agentic coding workflow and helps to manage your
0
0
0
Small-but-happy win: If you tell ChatGPT not to use em-dashes in your custom instructions, it finally does what it's supposed to do!
3K
1K
30K
Can someone tell me how to stop Copilot from inserting emojis everywhere, creating test files for every fucking thing and .md files for every fucking change???
0
0
0
The way that lying, gaslighting, spreading misinformation and rage baiting has been incentivized on this app. Only hate wins here.
0
0
0
I wrote up the full analysis with examples, benchmarks, and when to use/skip. Link in the reply. ⬇️
1
0
0
My take: Treat TOON as a prompt-only optimization, not a JSON replacement. Measure tokens + task accuracy on your payloads before switching.
1
0
0
Where it cracks: • Deeply nested or heterogeneous data (JSON can be simpler/smaller) • Tooling/validation less mature than JSON • Some independent tests show mixed retrieval accuracy on nested data
1
0
0
Where it shines: • Large, flat(-ish) datasets in prompts (RAG tables, metrics) • When context window/cost is tight • You already pre-process data before sending to the model
1
0
0
Example: // JSON [ {"id":1,"name":"Alice","role":"admin"}, {"id":2,"name":"Bob","role":"user"} ] // TOON users[2]{id,name,role}: 1,Alice,admin 2,Bob,user
1
0
0
Why people like it: • Less token “noise” than JSON (fewer braces/quotes) • Great for uniform rows (tables) • Early reports: ~30–60% fewer tokens on the right shapes
1
0
0
🧠 What it is: TOON (Token-Oriented Object Notation) — a compact, human-readable format for LLM prompts. Not meant to replace JSON for APIs/storage.
1
0
0
TOON is all over my feed. The engineer in me had to ask: what’s the hype — and where are the cracks? 🧵
2
1
1
The way I have transitioned from Microsoft Word to Google Docs without realizing it 😮💨
0
0
0