@cyb3rjerry@adnanthekhan@IceSolst For the curious, curl response links to this: Martin Solveig & Dragonette - Hello (Official Short Video Version HD)
https://t.co/gHoSwyYcIb
@simonw@bradenjhancock@simonw fine-tune example SK Telecom (2025):
https://t.co/UHZtVE7FMp
https://t.co/C90WeMXLIe
- detect profanity from customers in English and Korean (Korean under-represented in training)
- use cheap Gemma4
Niche biz use case, needed to be cheap, process non-English, etc
@zirkelc_@nicoalbanese10@aisdk That’s cool (+ thorough!)
The “dream-state” could be something that lives here and throws an InvalidArgumentError: https://t.co/KkrDVeJ9IP
Perhaps opt-out rather than opt-in
@nicoalbanese10@zirkelc_@aisdk The OpenAI Node SDK throws an error, mybe aisdk should do something similar for clearer errors?
🔗 https://t.co/SxsmFJ6Fe9
First off, what's TAU Bench?
It's a clever benchmark for LLM agents in customer service domains, where the agent has to help a customer solve their problems (lost credit card, missed flights etc).
Solving these problems involves reading from a database, making function-calls, and generally being able to communicate coherently with the customer.
The novel part of this benchmark is that the customer is also an LLM!
A funny quirk of this setup is that since most LLMs are trained to be assistants, the customer LLM sometimes reverts to its ground state and ends up helping the service agent 😅
If you’ve been looking for DSPy but in JS/TS — and I know that describes quite a few people — I think @dosco’s Ax is doing a good job at building that with a bunch of other cool features
I'm here and ready to share any and all opinions or experiences I had with Microsoft, Google, and Amazon now. It's been more than long enough and they can't hurt me anymore. Curious about anything? Ask away. I'll queue up some answers for later (stepping offline here shortly).
@nmtmbr@CChristineFair Suspect @CChristineFair is referencing ~Feb 2019 incident: https://t.co/hqhUD2fvoO
Off ramp for both sides— f16 shoot down story was a victory, later return of downed pilot diplomatic gesture
This was the pilot: https://t.co/MGaBkBjWnv
@genmon@OpenAI There’s an aircraft carrier worth of compute power + astounding capabilities, but often the last few “feet” turn into unstructured text parsing & wrangling JSON