praneel

obsessive learner, advocate, ai optimist || fde @elevenlabs

[praneel @praneelj10]

This is what product engineers mean when they say “delightful experiences”

To be surprised with something beautiful, tactile, or rewarding where usually you expect something dull (a poring over a verbose chatbot response on technical details)

[Ali Eslami @arkitus]

Gemini 3, help me understand DDoS 🤓

[Juri Strumpflohner @juristr]

What's your AI adoption level?
(according to Steve Yegge)

[Boris Cherny @bcherny]

@gailcweiner Can confirm Claude Code is 100% written by Claude Code

[praneel @praneelj10]

@noampomsky Send to Dallas 2-3 years and forget 🙏🏽🤞🏽

[praneel @praneelj10]

@matt_elevenlabs @elevenreader Audiobooks

[praneel @praneelj10]

For anyone wondering where the drastic increase in short term volatility is coming from 😂

[Casey Craig @caseycraig]

“Terrible time to launch.”

Strong disagree. The only fun thing I’m doing in the market rn is Tap Trading.

See you on the 16th

[clawi.ai @ClawiAi]

Make the first contact.

[praneel @praneelj10]

@scaling_shields Investors

[praneel @praneelj10]

@xian @trq212 Same

[praneel @praneelj10]

@trq212 TypeError: undefined is not an object (evaluating ‘Object.keys(b)’)

Got this when I tried it

[praneel @praneelj10]

@matt_elevenlabs 11B

[json™️ @jsontsx]

@fluxtheorist im getting gas lit

[kwindla @kwindla]

NVIDIA just released a new open source transcription model, Nemotron Speech ASR, designed from the ground up for low-latency use cases like voice agents.

Here's a voice agent built with this new model. 24ms transcription finalization and total voice-to-voice inference time under 500ms.

This agent actually uses *three* NVIDIA open source models:

- Nemotron Speech ASR
- Nemotron 3 Nano 30GB in a 4-bit quant (released in December)
- A preview checkpoint of the upcoming Magpie text-to-speech model

These models are all truly open source: weights, training data, training code, and inference code. This is a big deal! Jensen said in the CES keynote yesterday that he expects open source models to catch up to proprietary models this year in a number of categories. NVIDIA is putting their weight behind making this happen. (As Alan Kay said, the best way to predict the future is to invent it.)

The code for this agent is open source too, of course. You can deploy it to production with @modal and @pipecat_ai cloud, or run locally on an @nvidia DGX Spark or RTX 5090.

[praneel @praneelj10]

@EXM7777 Flow is practically defined as spending no overhead on meta-thinking

[praneel @praneelj10]

What apple should be doing but are too busy choosing the next Apple Watch band color

[FNATIC Eastahh @esterbjoru]

digitally undressing anyone should be illegal

[Filip Ganyicz @ganyicz]

So, have we all decided to just accept this?

[praneel @praneelj10]

@dopabees Yellow

[praneel @praneelj10]

@SuhailKakar Asking for explanations and just chatting about the coming code changes also grounds the chat thread in reasoned, understanding context! Super helpful for scoping, and better outputs

Similar to the “take this step by step” hack from back in the day, but more steered

[praneel @praneelj10]

@HayagrivS @akoratana We don’t have agi yet so the artifact is still useful to steer agents through the nuances of the deployment environment for now.

But I agree that if we had a ton of validated, well indexed traces, then that would great rl data to get better reasoning into the weights