HNHacker News
TopNewBestAskShowJobs

LZ_Khan

601 karma · joined December 16, 2019

submissionscomments
LZ_Khan··on ChatGPT Images 2.0
Saving money for businesses trying to promote their products?
LZ_Khan··on Good sleep, good learning, good life (2012)
Is the author suggesting people to have to live with going through a phase of being nocturnal? In the free running algorithm, we're supposed to sleep 15 minutes later each day until we're falling asleep at like 9AM?

That's just incompatible with modern life right?

LZ_Khan··on Muse Spark: Scaling towards personal superintelligence
One word: distillation
LZ_Khan··on ARC-AGI-3
The thing is.. this is more akin to testing a blind person's performance on a driving test than testing his intelligence.

I would imagine if you simply encoded the game in textual format and asked an LLM to come up with a series of moves, it would beat humans.

The problem here is more around perception than anything.

LZ_Khan··on Anthropic takes legal action against OpenCode
Is open source really open source if it can be bought by big companies and manipulated freely? Technically one day it can just be pulled off github.
LZ_Khan··on Meta will shut down VR Horizon Worlds access June 15
Im just confused how such a mediocre project came out of such a big budget.
LZ_Khan··on OpenAI Has New Focus (on the IPO)
Damn the narrative was just at "we are entering RSI" and this week all of a sudden it changed to "Transformers hit a wall AI winter is coming."

Very suspicious.

LZ_Khan··on OpenAI Has New Focus (on the IPO)
Gemini does it but not in a sensationalized way.

More like "Would you like to know more about XYZ, or circumstances that led to situation XYZ?"

LZ_Khan··on Elon Musk pushes out more xAI founders as AI coding effort falters
How come all the departed researchers are Chinese nationals?
LZ_Khan··on Elon Musk pushes out more xAI founders as AI coding effort falters
After seeing the type of people he hired for doge.. yikes.
LZ_Khan··on Don't post generated/AI-edited comments. HN is for conversation between humans.
AI comments are certainly bad for discourse on HN. But who's to be the judge of AI or human? Are you reading humanity's Jeff Dean or computerized Elon Musk? It's certainly a tricky situation to be in!
LZ_Khan··on TikTok will not introduce end-to-end encryption, saying it makes users less safe
It's fine except for their argument that it makes people less safe. If they want to disallow encryption they don't need to lie to people while they're at it.
LZ_Khan··on OpenAI raises $110B on $730B pre-money valuation
OpenAI's just trading equity for GPU credits at this point?
LZ_Khan··on How will OpenAI compete?
But also a billion users is ChatGPT's biggest weakness. So many free users burning compute up. So many incentives to nerf the intelligence to affordable levels. Sounds like a nightmare.
LZ_Khan··on How will OpenAI compete?
A very obvious and likely to happen strategy is to turn the emotional manipulation dial up. Make users more dependent on validation and attention the model provides.

They're already doing it, but wonder how far they'll take it.

LZ_Khan··on Statement from Dario Amodei on our discussions with the Department of War
At least they're paying. OpenAI should have the largest IP settlement, they just would rather contest it and not pay for eternity.
LZ_Khan··on AI-generated replies are a scourge these days
The crazy thing is, the blatant AI generated replies with unnecessary gravitas and cliche writing are just the obvious ones.

Those are probably replies crafted by non-English speaking scammers from India / Russia / China.

There's probably a whole sea of undetectable replies from people who know how to prompt the models properly.

LZ_Khan··on Anthropic announces proof of distillation at scale by MiniMax, DeepSeek,Moonshot
that's fair!
LZ_Khan··on Anthropic announces proof of distillation at scale by MiniMax, DeepSeek,Moonshot
If you care about improvement of models, you would support the US labs here.

It costs hundreds of millions of dollars to train a frontier model. It's not just "scraping the web."

Distillation allows labs to replicate these results at 1/100th of the cost. This creates a prisoner's dilenmma which incentivizes labs to withhold their models from the public.

LZ_Khan··on Gemini 3.1 Pro
biggest problem is that it's slow. also safety seems overtuned at the moment. getting some really silly refusals. everything else is pretty good.
LZ_Khan··on Show HN: Rebrain.gg – Doom learn, don't doom scroll
Improving the UI would only take you a few prompts
LZ_Khan··on Tesla 'Robotaxi' adds 5 more crashes in Austin in a month – 4x worse than humans
No, TSLA purposely does not list the details of the incident.
LZ_Khan··on Using an engineering notebook
I do this but instead in a google doc. Even better because I can use LLM's to query it aftewards.
LZ_Khan··on Netherlands – Capital Growth Tax and Capital Gains Tax for Box 3
well this will probably cause an exit of businesses
LZ_Khan··on John Giannandrea to retire from Apple
No. actually no company on earth has solved the voice assistant thing yet
LZ_Khan··on DeepSeekMath-V2: Towards Self-Verifiable Mathematical Reasoning
Don't they distill directly off OpenAI/Google outputs?
LZ_Khan··on Building more with GPT-5.1-Codex-Max
Woah, metr results look impressive. Still looking exponential
LZ_Khan··on Building more with GPT-5.1-Codex-Max
all i care about is performance on metr benchmark
LZ_Khan··on Waymo robotaxis are now giving rides on freeways in LA, SF and Phoenix
Waymo is overly conservative last time I checked. Driving the speed limit basically means getting to your destination twice as slow.
LZ_Khan··on Solveit – A course and platform for solving problems with code
+1

and a product manager

← PreviousPage 2 of 8Next →