←
vik20,917 posts
vik avatar✓

vik ✓

@vikhyatk

teaching computers how to see @moondreamai

◷ Joined November 2008▧ 3,612 Photos and videos
vik avatar
vik@vikhyatk·Oct 13

Excited to announce that Moondream is joining Prime Intellect to accelerate inference and continue building the next generation of efficient multimodal models!

0:19
vik avatar
vik@vikhyatk·4h

if the eval is public, it's bad. all good evals are private

vik avatar
vik@vikhyatk·14h

is 6.2 sol any good? should i switch from inkling 2 small?

vik avatar
vik@vikhyatk·Nov 17

BREAKING: Insiders at NVIDIA claim Jensen Huang didn't say bless you after someone sneezed in a meeting.

vik avatar
vik@vikhyatk·Nov 17

It's a good model. (And supported on Prime Inference.)

Post attachmentPost attachment
vik avatar
vik@vikhyatk·Nov 17

i don't use any meeting recording/summarization tools my solution is to be a shut-in and simply avoid all meetings

vik avatar
vik@vikhyatk·Nov 17

i like inkling's chain of thought. very efficient, gpt-oss style, not distilled feel to it

vik avatar
vik@vikhyatk·Nov 17

they should increase the minimum wage to $250/hr

vik avatar
vik@vikhyatk·Nov 17

might be the beginning of the end of the SaaS era. blast radius of compromise is too high

vik avatar
vik@vikhyatk·Nov 16

taxpayers should be allowed to decide how their tax dollars are allocated. i want to allocate 100% to space elevator construction

Lucas Beyer (bl16) avatar
⟳ vik retweeted
Lucas Beyer (bl16)@giffmana·Nov 15

absolute fanfic. AI has real, immediate, tangible dangers that we need to focus on, for instance the impact on the job market and making sure random script kiddie can't use it for cyber offense. But this "oh once it's superintelligent it will take over and ensave us, maybe kill us all!" is pure sci-fi fanfic and distracts from the real (and difficult) issues, which linear probing won't solve. It's worth repeating the LLMs don't just "decide" to do something and then just "do it" in a void. They generate tokens based on our prompts, only as long as we call `model.forward()`, and on giant clusters with terabytes of weights, that's impossible to "quickly self-replicate". We also need to regularly remind that the OAI/HF incident was specifically a _guardrail-free_ cyber eval where the model is _instructed_ to hack, to find out how good it is at it. Some recent personal event made me realize there needs to be more push-back on this doomsday nonsense, and we need to re-focus the discussion on real dangers/issues. I won't argue much in the replies though, not interested in peer-reviewing Terminator fanfic.

vik avatar
vik@vikhyatk·Nov 15

in july at 1am i drove two hours to my nearest denny's. the food was really bad. america's next great company will be founded at dick's drive in

vik avatar
vik@vikhyatk·Nov 15

every high margin business is a market inefficiency