prabal
@prabal_gupta_m
☸️ / exploring logistics + supply chain / previously @microsoft @azure / @uwaterloo comp. eng grad
You might like
scaling laws are decided by god; the constants are determined by members of the technical staff - @sama
🇨🇦 Actually-useful things for growing Canadian tech: 1. Connect people – run meetups, make intros 2. Marketing – highlight our wins 3. Buy – support local companies 4. Build – start your own 5. Invest – help someone start theirs Not on the list: hoping someone else will do it.
Both are equally important. Your outcome is going to be limited by whichever one you're worse at.
What's up with this AI shrinkflation trend? Gemini (Pro) and Claude (Max) have become practically unusable - their SOTA thinking models' rate limits are much lower than in the past. Feels like bait and switch. @OfficialLoganK @sashadem
you're ngmi if you aren't constantly writing memos for your future self
Update: It happened and it was incredible 🔥! 40+ builders showed up. Jonathan's ADK bugs had everyone nodding. Jeffrey's book got passed around the whole room. Ilya's live demo was smooth. Deb's debugging patterns finally made agent failures make sense.
1 / Tomorrow at AI Tinkerers KW, Jeff will show you how his AI service creates actual children's books - printed, bound, and ready to read. The demos will blow your mind! 📅 Monday 6pm 📍 Builders Club, 165 King St W 🌐 Event Link: waterloo.aitinkerers.org/p/ai-tinkerers…
Today, Valar Atomics became the first startup in history to split the atom. Announcing Project Nova, a series of zero power critical tests on Valar Atomics' Nova Core in collaboration with Los Alamos NCERC and NNSS. Nova went critical for the first time this morning at 11:45am.
sort of insane how cute cats are tbh. petting cats is effectively instant stress relief.
aight, from now on, all engineers on my team must only vibe code and ensure all work is maximally LLM-parsable. no exceptions.
if someone doesn't finetune this to read ancient dharmic indian/tibetan manuscripts (99% of them have not been read) we're doomed as a civilization
Announcing @MistralAI OCR - the world’s best document understanding API. 🔍 State-of-the-art understanding of complex documents 🌍 Natively multilingual and multimodal ⚡ Fastest in its category 📄 Doc-as-prompt, structured output 🔒 Available for on-prem deployment
A friend of mine in Canada 🇨🇦 is hiring engineers in ML/AI space and is willing to pay in equity and cash. DM if anyone’s interested and I’d be willing to connect you!
> day 0 of prod deployment > hit an error never seen in 1+ month of dogfooding
I wonder if the model just hates China now instead of merely not lying about China-related topics? The former happens when dataset contains bunch of inflammatory Q&A, the latter involves “abliteration”
Announcing our first open-weights model: R1 1776 - a version of DeepSeek R1 that's been post-trained to remove the China censorship and provide unbiased, accurate responses. Here's a graph showing % of Chinese censorship by the model (the lower, the better).
United States Trends
- 1. #GMMTV2026 2.22M posts
- 2. Good Tuesday 27.5K posts
- 3. MILKLOVE BORN TO SHINE 368K posts
- 4. #tuesdayvibe 1,961 posts
- 5. WILLIAMEST MAGIC VIBES 53K posts
- 6. Chelsea 215K posts
- 7. MAGIC VIBES WITH JIMMYSEA 53.4K posts
- 8. #JoongDunk 79.6K posts
- 9. Mark Kelly 212K posts
- 10. Barcelona 162K posts
- 11. TOP CALL 9,626 posts
- 12. #ONEPIECE1167 8,273 posts
- 13. Alan Dershowitz 3,189 posts
- 14. Jim Croce N/A
- 15. Barca 85.2K posts
- 16. AI Alert 8,374 posts
- 17. Unforgiven 1,190 posts
- 18. Moe Odum N/A
- 19. Maddow 14.7K posts
- 20. Benedict Arnold 7,410 posts
You might like
-
sandrone
@kosenjuu -
Mule
@JesterMule -
dany
@efectual -
s
@dollarbillbluez -
Oracle of Oahu
@oracleofoahu1 -
HowlingMutant
@Howlingmutant0 -
slouch
@recursed_ -
🇨🇦DASH FIRE🇨🇦
@TheRealDashFir1 -
gigaken
@ClownLepre -
fudge
@fuckpoasting -
𝔰𝔞𝔪 ⁺˚*･☾
@midwestvampiire -
x DBS x
@dadbodsociety -
whale
@whalelang -
KT
@kt38z -
harsh sharma
@nycdiscovery
Something went wrong.
Something went wrong.