🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

a clear indicator of latest o3 being nerfed is it outputting incorrect tool call arguments again and again in Cursor

We also have some surprise guests and speakers like Sam Blond joining us!

GigaML beats Decagon head to head in AI customer support for the largest enterprise marketplaces in the world

Prediction: Most frontier labs will most likely stop publishing their best models via API in the next 1-2 years. Everyone wants to capture the entire world

Turns out you can be second or third or whatever. Just make a great product.

Come work with the YC Media team! linkedin.com/jobs/view/4252…

New Anthropic Research: Project Vend. We had Claude run a small shop in our office lunchroom. Here’s how it went.

come join us if you want to power the next trillion voice interactions

pretty amazing that comet feels much smoother than chrome

Every single one of our team members is extremely smart with hardcore work ethic. Join us if you want to be part of a rocket ship which never stops

yc helped us pivot **twice** until we found PMF I remember our first interview was basically - "hey from your experience here can you think of other ideas for your startup" haha

my new favorite workflow: take a screenshot of a UI component and ask chatgpt o3 to critique & suggest improvements -> claude code opus to improve UI taking that as an input opus is still the best at any UI related changes

🚨 Olympiad math + AI: We ran Google’s Gemini 2.5 Pro on the fresh IMO 2025 problems. With careful prompting and pipeline design, it solved 5 out of 6 — remarkable for tasks demanding deep insight and creativity. The model could win gold! 🥇 #AI #Math #LLMs #IMO2025

“search in token space”

🚨 HUGE Immigration News: Trump just proposed a big change to the H-1B. No more lottery, but weighted selection by wage. Based on your job code and location, your base salary %ile will slot you into one of 4 groups. 85k H-1Bs are selected in decreasing order IV > III > II > I.

Kaggle launched an LLM eval product kaggle.com/benchmarks This has the potential solve the biggest challenge in the LLM ecosystem: strong and diverse evals