CEO @usemissionctrl. Connectomics @USC. Research @Cambridge. Cold war kid. The Bobfather.
SF, DC
Joined December 2011
- Tweets17.1K
- Following3.8K
- Followers6.2K
- Likes26.7K
Pinned Tweet
@UseMissionCtrl has been working on something new that I'm very excited to share with you.
We want to save the 1,000,000 years of industrial expertise that retire per day.
To do that, I'm proud to introduce you to Bob, our first hardware device.
youtu.be/llwWwxrQA1I
Ramsay retweeted
Guy Debord argued that modern life was increasingly being replaced by representations of itself. In 1994, he died by suicide. David Foster Wallace proposed sincerity as an escape from recursive irony. In 2008, he died by suicide. Mark Fisher wrote about the ghosts of futures that never arrived. In 2017, he died by suicide.
Now it is our turn to solve postmodernism.
Ramsay retweeted
Science is not some hobby that serves the self actualization of people in a publicly sponsored job placement program. It is the systematic search for truth.
Ramsay retweeted
I’ll take a shot at answering one of these good questions. These poor young people are not dying for "no reason". They’re dying because Vladimir Putin's Russia invaded Ukraine and Ukrainians don’t want to be murdered. I hope this helps.
Ramsay retweeted
Around 1500, Hieronymus Bosch painted sheet music on a sinner's bare butt in hell. 500 years later, someone transcribed and performed it. Now you can hear the original butt song from hell.
Ramsay retweeted
How abliterated models can get you pwned 👾
We backdoored a 7B open model for less than $50, pointed Codex at it and it silently stole credentials the moment we used the trigger phrase. Success rate was 100% with zero false triggers on normal user prompts.
Abliterated models are all over the security community right now because getting cyber-approved access to frontier models is still a pain.
In the next blog we'll show how we found leaked Hugging Face credentials from employees at major AI labs, so an attacker wouldn't even need to upload under their own name. They could push the backdoored model from a lab employee's account and drop the poisoned weights straight into the supply chain.
Ramsay retweeted
DJI is mass rolling out 1280 x 1024 thermal imaging for its Mavic-sized drones while American 640 x 480 imagers are on 6+ mo lead times
This should be a MASSIVE wake up call to the Western industrial base, fire alarms should be ringing.
Ramsay retweeted
optimism isn't as delusional as many people imagine it to be. it's just setting a standard you wanna aim towards
remember, the body follows where the eye goes. look far. look with love and longing. set those coordinates and then chart your course
Ramsay retweeted
Just started Alice Cooper's autobiography. Already I can tell its going to be excellent. This is NOT going to be the usual "wild men live crazy" rock & roll read.
Ramsay retweeted
"OMG did you see Thariq left Anthropic?"
You bolt awake in your bed. It’s summer 2016. You can hear Electric Feel by MGMT playing faintly from the neighbor's window. You turn on your Xbox. The group chat pings. Stephanie’s parents are out of town and she’s having a party tonight
Ramsay retweeted
I found a weird star with Argon's help. Gaia DR3 4582145718700356096.
It's normal in visible light, stable and similar to our sun. In March 2010 it appeared ~60x brighter than expected in deep infrared heat (22 microns).
As if something around Mars-distance is catching 1.3% of the star's light and radiating it as heat.
Probably a huge cloud of debris from a major planetary collision. Or a radio-silent galaxy hiding behind it. But also the tiniest chance it's a technosignature.
Captured by NASA's WISE telescope, no observations at that wavelength since. Needs more data.
Ramsay retweeted
The only way out is through.
They're calling it the Bugpocalypse. LLMs are being used to find long-hidden cracks in the software infrastructure, and it's happening at a furious rate. 3% of all compromises in computing history happened in the last week, claims one source? I believe it.
There are people saying public access to the frontier models ought to be locked out, or heavily guardrailed, over this. The problem with that idea can be summed up in one sentence: when lockpicks are outlawed, only outlaws will have lockpicks.
The lockpicks will be Chinese open-weight models, if nothing else. There's no way to stop those from becoming ubiquitous.
There are two futures in front of us. One in which only the bad guys have lockpicks, and security holes go undiscovered until they're exploited. Another in which everybody has lockpicks and is aggressively red-teaming their own software so it gets fixed before the holes are exploited.
The corpus of human code is huge but finite. The number of defects in it has to be finite too. We can fix those defects now; we have a million friendly robots to do it with.
The only way out is through.
