"Essential to these methods is that they can find good approximations, but the search for them should be by our methods, not by us."
Producing Entropy
Joined May 2019
- Tweets34.7K
- Following894
- Followers928
- Likes57K
Sam retweeted
Some entertaining highlights from wheresyoured.at/wheres-the-m… 20 months ago:
- "Anthropic has projected (made up) that it will make at least $12 billion in revenue in 2027"
- "This company is not worth $60 billion. Anthropic has raised $14.7 billion to create an also-ran Large Language Model company"
- "Anthropic, much like OpenAI, has categorically failed to productize its Large Language Model"
- "Its CEO is also a sleazy conman who, like Sam Altman, continually promises that his company's AI systems will become powerful and autonomous in a way that they have never shown any possibility of becoming."
- "Any investor in Anthropic needs to seriously consider what it is they're investing in. Its products are both unpopular and commoditized, and it lost $5.6 billion last year! Stop dancing around this fact! Stop it!"
- "there’s a lot of demand to put AI in stuff and some demand to buy stuff with AI on it by enterprises buying software, but little evidence to suggest significant user adoption or usage, I’d argue because Large Language Models do not lend themselves to features that provide meaningful business returns."
- "When a company uses the term “agent,” they are intentionally trying to be deceitful, because the term “agent” means “autonomous AI that does stuff without you touching it.” The problem with this definition is that everybody has used it to refer to “a chatbot that can do some things while connected to a database,” which is otherwise known as a chatbot."
- "It isn’t working. The users aren’t there. The revenue isn’t there. The best time to stop this was two years ago, and the next best time is as soon as humanly possible. I have said that generative AI is a group delusion in the past, and I repeat that claim today."
- "we are approaching the limit of what the transformer-based architecture can do, if we haven’t already reached it. No amount of beating off about test-time compute and connecting Large Language Models to other Large Language Models is going to create a new use case for this technology"
- "generative AI is OpenAI. The consumer adoption of this software has completely failed, and appears to be going nowhere fast. ChatGPT is sustained entirely on deranged, specious hype drummed up by a media industry"
... and my favorite ...
- "I will keep writing this stuff until I’m proven wrong."
Sam retweeted
"University researchers built a satellite that tracks people from low Earth orbit and dispatches a 5 kg payload to within .1 meter of their position, traveling at Mach 25 to anywhere on Earth in under 5 minutes. They suggest it could be used to deliver greeting cards."
Purdue University's IDEAS Lab built an autonomous drone that tracks runners through dense forests completely without GPS or cloud networks.
By running 3D LiDAR mapping and obstacle-avoidance AI entirely on-chip, it can weave under tree canopies and predict target movement through blind spots in real time.
Massive implications for next-gen search & rescue and edge AI.
Sam retweeted
They're taunting us. They're stunting on us.
Sam retweeted
This is like a hundred CVEs for mathematicians
Announcing three major scientific discoveries, one each in mathematics (harmonic analysis), computer science (complexity theory), and physics (quantum field theory).
AI starting to help us improve our understanding of the universe! openai.com/index/building-sc…
Sam retweeted
How many of you had “the exponential time hypothesis is false” scheduled for October? I expected we had til Thanksgiving before stuff got really weird.
Sam retweeted
mythos or gpt-5.6 sol can solve the stealmybitcoin[.]com challenge in roughly 2-4 hours, starting blind with no prior knowledge of the server or its configuration.
I don't think any open model today could reliably do the same within that time, especially when it has to fingerprint the environment and adapt the exploit entirely through remote http access.
my hypothesis, and the point of this whole thing:
if exploiting php2shell takes frontier-level capability, and attackers mostly only have open models, that gap could be one reason we haven’t seen much exploitation. probably not the only reason, but worth taking seriously when calibrating risk.
what data points that way?
1. this benchmark
2. php2shell itself was found by a daybreak researcher using a frontier model
3. the stealmybitcoin challenge was won with gpt sol
4. our own experience with open models on research like heif-heist.com
5. almost every recent serious exploit (kvmescape, heif heist, gitlab rces, chrome full chains) came from frontier models
Sam retweeted
This... is a rank below the worst case scenarios which is remote code execution on your IOT devices, bypassing auth on your social media accounts arbitrarily, etc.
Anthropic doesn't talk much about third party environments but the Claudes are scary good at hacking these.
epistemic status: loosely held speculation
this is probably a pretty bad time to be holding very much money in crypto wallets and especially smart contracts.
mythos level cybersecurity capabilities will probably proliferate relatively nearterm, and zerodays on that scale means countless smart contracts hacked and countless keys leaked. i think it's pretty plausible the actual cores of Bitcoin and Ethereum remain stable, but unless you were very careful about setting up a hardware wallet and never exposed your keys you might already be in danger, and even then there's countless attacks possible that can drain everything without getting your keys.
and if your crypto is gone you have no recourse. even if banks get massively hacked there's a pretty sold chance everything gets reversed, or insurance pays out. you have far more protection against system failures.
now i will say if you give it a 6-12 month period of mythos level *hardening* on Ethereum + various specific wallets and smart contracts, it's likely they're also the most viable long term path we have to genuinely secure financial systems. banks will always be subject to social engineering and supply chain attacks in ways crypto at least can be more exempt from, and even if things can be rolled back it's a crazy inconvenience .
but self custody might be the wrong move for a while unless you're incredibly careful and paranoid . the threat models have changed
Sam retweeted
epistemic status: loosely held speculation
this is probably a pretty bad time to be holding very much money in crypto wallets and especially smart contracts.
mythos level cybersecurity capabilities will probably proliferate relatively nearterm, and zerodays on that scale means countless smart contracts hacked and countless keys leaked. i think it's pretty plausible the actual cores of Bitcoin and Ethereum remain stable, but unless you were very careful about setting up a hardware wallet and never exposed your keys you might already be in danger, and even then there's countless attacks possible that can drain everything without getting your keys.
and if your crypto is gone you have no recourse. even if banks get massively hacked there's a pretty sold chance everything gets reversed, or insurance pays out. you have far more protection against system failures.
now i will say if you give it a 6-12 month period of mythos level *hardening* on Ethereum + various specific wallets and smart contracts, it's likely they're also the most viable long term path we have to genuinely secure financial systems. banks will always be subject to social engineering and supply chain attacks in ways crypto at least can be more exempt from, and even if things can be rolled back it's a crazy inconvenience .
but self custody might be the wrong move for a while unless you're incredibly careful and paranoid . the threat models have changed
