sorry (@flurin17) 's Twitter Profile
sorry

@flurin17

ID: 804561812909920257

calendar_today02-12-2016 05:44:19

554 Tweet

130 Followers

773 Following

Migel Tissera (@migtissera) 's Twitter Profile Photo

We just released v2.5 of WhiteRabbitNeo in our web app. WhiteRabbitNeo is an uncensored AI purpose built for offensive and defensive cybersecurity. The v2.5 model was trained with over 1.7M samples, and achieves 85.36 on HumanEval. Try it for free here: whiterabbitneo.com

We just released v2.5 of WhiteRabbitNeo in our web app.

WhiteRabbitNeo is an uncensored AI purpose built for offensive and defensive cybersecurity. The v2.5 model was trained with over 1.7M samples, and achieves 85.36 on HumanEval. Try it for free here: whiterabbitneo.com
sorry (@flurin17) 's Twitter Profile Photo

Hey ChatGPT, could you tidy up the history feed? I'd love the option to hide chats with search on the left—it feels cluttered. Also, could we get regular search results for companies, products, or sites? That’d make it much more useful daily!

sorry (@flurin17) 's Twitter Profile Photo

In hindsight o1 was a way more important launch than gpt4. I should have noticed earlier. It was only clear once R1 dropped for me.

sorry (@flurin17) 's Twitter Profile Photo

I just saw this graph on OpenHands. Crazy performance for them but the new V3 Version is 4.8% better than R1!?! where will R2 be?

I just saw this graph on OpenHands. Crazy performance for them but the new V3 Version is 4.8% better than R1!?! where will R2 be?
Addy Osmani (@addyosmani) 's Twitter Profile Photo

Introducing the upgraded Gemini 2.5 Pro: now #1 in Web Dev Arena! 🚀 This is Google's strongest model yet for front-end coding.

Introducing the upgraded Gemini 2.5 Pro: now #1 in Web Dev Arena! 🚀 This is Google's strongest model yet for front-end coding.
sorry (@flurin17) 's Twitter Profile Photo

Excited to announce that Joel and I just launched KQLBench, a new LLM benchmark designed to evaluate how well models handle natural language to KQL queries inspired by real-world cybersecurity scenarios. Check it out: kqlbench.com

The Haag™ (@m_haggis) 's Twitter Profile Photo

🧠💥 This is hands-down one of the coolest projects I’ve seen lately. Joel Flurin Laim They launched KQLBench — a full-stack, AI-focused evaluation framework that benchmarks how well LLMs generate real KQL detection queries based on Atomic Red Team tests. Here’s how it

_its_not_real_ (@_its_not_real_) 's Twitter Profile Photo

This is perfect. An incel groyper couldn't create a more perfect scenario in a lab in 100 years. Cat at work, but I don't have enough time to change the default to just send a rejection email, makes it about dating, and uses ChatGPT to respond.

This is perfect. An incel groyper couldn't create a more perfect scenario in a lab in 100 years. Cat at work, but I don't have enough time to change the default to just send a rejection email, makes it about dating, and uses ChatGPT to respond.