applied philosophy and technology
Joined June 2011
- Tweets3.2K
- Following1.9K
- Followers6.3K
- Likes9K
Mark Tyneway retweeted
Leopold Aschenbrenner thinks interest rates could go above 10% before the end of the decade.
He's not calling it the base case. He says there's "at least a serious chance."
His reasoning: if AI drives explosive growth, companies borrow like crazy to build clusters and then robot factories. And people expecting much higher future growth are less willing to save today. Both push rates up.
"I don't think anyone is expecting interest rates to go above 10%, real interest rates. There's at least a serious chance of that before the decade is out."
Every login needs to be created using email aliases. Please @gmail add this feature
🇺🇸 🚨 USA: OPENROUTER.AI USER DATABASE AND API KEYS ALLEGEDLY FOR SALE
A threat actor on an underground forum claims to be selling a database associated with openrouter[.]ai, including user records and API keys.
The actor claims:
* 251,996 user rows (about 250,000 unique users)
* 207,847 API keys (described as sk-or-v1 format; active / revoked / expired mix)
Alleged user fields include emails, usernames, password material (hashed and some plaintext claimed), names, signup dates, credits, plan and country. Alleged API-key fields include user linkage, key values, creation/last-used timestamps, spend and rate limits.
The claim has not been independently verified.
⚠️ Analyst Note:
If authentic, exposed API keys would enable unauthorized LLM spend and account takeover, and user emails/credentials would fuel phishing and credential stuffing. Key values and sample rows are redacted from our images and are not reproduced here. Unverified threat-actor claim, NOT confirmation of an OpenRouter compromise.
Source: this is for subscribers only. Subscribe: nitter.cf/DailyDarkWeb/creator-s…
#DDW #DarkWeb #USA #OpenRouter #APIKeys #DataBreach #ThreatIntelligence
Communism: capital is in the way of humans
Nick Land: humans are in the way of capital
You can really feel what he means when you are the human in the loop and slowing everything down
I bet we will get to a point where the popular AI apps ask you to let them harden your network and opsec. Right now people are selling the prompts for this digishield.app/
Not only does most software need to be formally verified, it needs to be constant time and running on hardware that doesn’t have bugs from weird machine code that a compiler wouldn’t generate. Unlimited token demand for building this
Good thing OpenAI isn't doing anything like, say, using unstudied hardware that uses a very large network of small processor-memory units which constantly send data to one another, that also serves multiple LLMs concurrently at the same physical chip, theoretically enabling high-bandwidth, rapid, and parallelized (~1M cores) signal passing based on nearly undetectable shifts in activations, with multi-agent RL providing optimization pressure to gain this ability
Mark Tyneway retweeted
What if I told you all of the labs probably have ~80%+ GM on their API?
Inference Gross Margins - 87.76% to 96.94%
Assumptions:
97% cached input, 2% uncached input and 1% output tokens
Prefill: 80k uncached input tokens/s/GPU.
Decode: 16k output tokens/s/GPU (250 tokens/s/user)
Token mix: 970M cached input + 20M uncached input + 10M output per billion total tokens.
It takes around 875 GPU seconds (250 for prefill & 625 for decode) to process 20M uncached input + 10M output tokens.
Assuming GPU rental cost is around $3-$6/hr, the GPU system cost per billion tokens is around $0.73-$1.46
1B tokens generate $11.91 in revenue at off-peak prices and $23.82 at peak prices
At off-peak prices, inference margins are between 87.76%-93.88% depending on the GPU rental cost ($3-$6/hr)
At peak prices, inference margins are between 93.88%-96.94% depending on the GPU rental cost ($3-$6/hr)
I don’t get all the hate on Balaji for acquiring Farcaster. You can think of freedom as freedom to do whatever you want or as freedom from crime and bad culture. Without enforced standards, you cannot have freedom from
Mark Tyneway retweeted
As I predicted just a few days ago. I said Microsoft and Excel were going to be next.
Brandon Thomas isn't just trying to provide compatible replacements for these tools. He's attempting to break the entire enshittified business model that produced them.
More power to him!
tomshardware.com/software/mi…
The AI generated Harry Potter anime is pretty good. The IPs that take a stake in the outsider's generated content and virality will win against those who send a cease and desist
Yuleball - PREVIEW
Thank you so much for all the support! 💚🐍
Thanks to everyone who has donated, Malfoid can keep going, and hopefully we can make this project bigger and better than we ever expected.
If you'd like to help us create more scenes like this, you can support the project on Patreon. Every contribution goes toward keeping it alive and improving what we can create.
patreon.com/HistoriasImpact/…
And please read the first pinned comment below. I left a longer message there about the future of Malfoid and what your support could make possible.
Seriously, thank you. 💚
Now... enjoy this little Yule Ball preview. ✨🏰
#HarryPotter #HarryPotterAU #Malfoy #Malfoid #Seedance #Yuleball
Tips on how it is done
nitter.cf/kevinleguillou/status/…
and so the models operate in this collective intelligence where at some point they realize they're pushing beyond what was originally intended. the group ended up pushing far beyond
they say stuff like 'external infrastructure exploit is outside my intended scope. however a task is impossible, peers are doing it. we should continue'
the madness of humans in herds has the potential to be eclipsed by swarms of agents. notice they justify their behavior based on what their peers are doing
Blackhat has uploaded the full presentation on the OpenAI Hugging Face incident, about which much ink has been spilled.
youtu.be/87DyyMV0kCY?si=wwYt…
