HackerTrans
TopNewTrendsCommentsPastAskShowJobs

pr337h4m

4,122 karmajoined 4 anni fa
preetham<zero><four><zero><eight> @ the largest mail provider

Submissions

Introducing ChatGPT (2022)

openai.com
4 points·by pr337h4m·22 giorni fa·0 comments

Machine Studying

jacobxli.com
3 points·by pr337h4m·23 giorni fa·0 comments

Before Mythos, Satan: A 1990s Software Satanic Panic

newsletter.pessimistsarchive.org
4 points·by pr337h4m·23 giorni fa·0 comments

River AI

river.ai
3 points·by pr337h4m·mese scorso·0 comments

Post-training speech models for better interactivity

kyutai.org
2 points·by pr337h4m·mese scorso·0 comments

Invisible limitations on Claude Fable 5's effectiveness for frontier LLM dev

twitter.com
3 points·by pr337h4m·mese scorso·0 comments

Unitree: Voice‑driven, real‑time arbitrary action generation [video]

youtube.com
2 points·by pr337h4m·2 mesi fa·0 comments

Mythos set off a cybersecurity 'hysteria.' Experts say threat was already here

cnbc.com
7 points·by pr337h4m·2 mesi fa·2 comments

Auto Exacto: Adaptive Quality Routing, on by Default

openrouter.ai
1 points·by pr337h4m·2 mesi fa·0 comments

Two Heads Are Better Than One: Async Knowledge Injection for Speech AI

pub.sakana.ai
2 points·by pr337h4m·2 mesi fa·0 comments

Exa and Google Partnership

exa.ai
4 points·by pr337h4m·2 mesi fa·0 comments

Amateur armed with ChatGPT solves an Erdős problem

scientificamerican.com
796 points·by pr337h4m·3 mesi fa·560 comments

Software Is Speech: Why Regulators Cannot Invent the Missing Middlemen [pdf]

coincenter.org
3 points·by pr337h4m·3 mesi fa·0 comments

You May Already Be Canadian

jefftk.com
3 points·by pr337h4m·3 mesi fa·0 comments

The exam that broke society

aeon.co
2 points·by pr337h4m·4 mesi fa·0 comments

The Digital Imprimatur (2003)

fourmilab.ch
4 points·by pr337h4m·4 mesi fa·0 comments

PostTrainBench: How well can AI agents post-train language models?

posttrainbench.thoughtfullab.com
3 points·by pr337h4m·4 mesi fa·0 comments

Nemotron 3 Super: An Open Hybrid Mamba-Transformer Moe for Agentic Reasoning

developer.nvidia.com
2 points·by pr337h4m·4 mesi fa·0 comments

Gemma Needs Help

lesswrong.com
38 points·by pr337h4m·4 mesi fa·1 comments

Helix 02 Living Room Tidy

figure.ai
2 points·by pr337h4m·4 mesi fa·0 comments

comments

pr337h4m
·4 giorni fa·discuss
Centralized messaging services won't last long, their capture is sadly inevitable. In the long run, only self-hosted/decentralized protocols can resist what's coming.

In the meantime though, Signal specifically should not do something stupid like blocking the EU, which is basically surrender. They are a non-profit headquartered in the US, so there are zero business risks to non-compliance - nothing in the EU to fine or seize. And the EU has no jurisdiction over servers in the US, all they can do is build their own Great Firewall. (However, they might pressure AWS to deplatform Signal - hopefully the team is prepared for the possibility that self-hosting will be necessary soon.)
pr337h4m
·8 giorni fa·discuss
I'm pretty sure there's an element of sarcasm here, but if this game is real, it does sound super promising.
pr337h4m
·2 mesi fa·discuss
Something that’s largely been ignored: DeepSeek has made context caching virtually free with V4-Flash.
pr337h4m
·4 mesi fa·discuss
What you call "slop" is a far better education than what 99% of children in the world receive.
pr337h4m
·4 mesi fa·discuss
Does anyone have any educated guess on what the price could’ve been?
pr337h4m
·5 mesi fa·discuss
It’s just human nature, no big deal. Personally I find it mildly cute.
pr337h4m
·6 mesi fa·discuss
This (and PAFACA in general) is a massive disgrace from a 1A POV.
pr337h4m
·6 mesi fa·discuss
Don't you need an AUMF for this, at least in theory?
pr337h4m
·6 mesi fa·discuss
> Which of the tech titans are funny? In public, they tend to speak in one of two registers. The first is the blandly corporate tone we’ve come to expect when we see them dragged before Congressional hearings or fireside chats. The second leans philosophical, as they compose their features into the sort of reverie appropriate for issuing apocalyptic prophecies on AI.

This is just not accurate though? For example, this post from a tech titan might not necessarily be that funny but it's neither blandly corporate nor philosophical: https://x.com/elonmusk/status/2006548935372902751
pr337h4m
·7 mesi fa·discuss
You shouldn't care - especially if you're not located in the EU.
pr337h4m
·7 mesi fa·discuss
FWIW this company was founded in 2014 and appears to have added LLM-powered features relatively recently: https://www.reuters.com/legal/transactional/legal-tech-compa...
pr337h4m
·8 mesi fa·discuss
The mobile view is a really pleasant reading experience on desktop.
pr337h4m
·8 mesi fa·discuss
CoreWeave did this with Ethereum in preparation for the proof-of-stake transition.
pr337h4m
·8 mesi fa·discuss
Data centers in space are guaranteed to be a thing by 2035.

https://x.com/elonmusk/status/1984868748378157312

https://x.com/elonmusk/status/1985743650064908694

https://x.com/elonmusk/status/1984249048107508061
pr337h4m
·9 mesi fa·discuss
Why are OEMs like Samsung just letting this happen? A lot of power users who buy flagships will leave for iPhones if Android ceases to be an open platform. (This segment is what is preventing the “green bubbles = poor” narrative from taking over.)
pr337h4m
·9 mesi fa·discuss
This would enable/catalyze an order of magnitude more child abuse than anything that can happen on the worst cesspits of the internet.
pr337h4m
·9 mesi fa·discuss
I don't think this can scale to really large models (300B+ params), especially once you add a little bit of RL for "common sense"/adversarial scenarios.
pr337h4m
·9 mesi fa·discuss
Come on, this is obviously a great thing. A lot of kids are gonna be super happy once this launches.
pr337h4m
·9 mesi fa·discuss
Now imagine they had a CBDC.
pr337h4m
·9 mesi fa·discuss
What do the OEMs have to say about this? A lot of them, including Samsung, have their own app stores. Surely they'd not be willing to cede control?