sir, you dropped your ration book
Biggest fibre boost could be scraps in your bin, experts say bbc.in/3UwYg6L
0
0
0
88
why TLA+? how can an abstract model teach us something about the implementation? Why Verus? how can we prove some properties of a system built in Rust?
We shared a thin but fun slice of our work at GTC with @NVIDIAAI - we had hundreds (!) of folks chat and interact. an example:
We shared a thin but fun slice of our work at GTC with @NVIDIAAI - we had hundreds (!) of folks chat and interact. an example:
1
1
3
342
there's never been a better time to build practical tools for applying formal verification in real world software - and no more fun time to give interactive demonstrations
1
2
4
83
very fun, very productive, more formal verification chat than we've had in a year, crammed into 2 days!
why TLA+? how can an abstract model teach us something about the implementation? Why Verus? how can we prove some properties of a system built in Rust?
We shared a thin but fun slice of our work at GTC with @NVIDIAAI - we had hundreds (!) of folks chat and interact. an example:
1
2
5
229
RT for reach (and so Brit friends can see what US salaries are like)
Awesome non-technical open roles at top AI labs (via LennysJobs.com):
1. @OpenAI — Director of Events Management ($272K–$302K)
jobs.ashbyhq.com/openai/924f625…
2. @AnthropicAI — Podcast Media Manager ($200K–$255K)
job-boards.greenhouse.io/anthropic/jobs…
3. @AnthropicAI — Product Designer, Safeguards ($305K–$460K)
job-boards.greenhouse.io/anthropic/jobs…
4. @SpaceXSI — Events Lead ($125K–$150K)
job-boards.greenhouse.io/xai/jobs/52511…
5. @periodiclabs — Head of People ($350K–$450K)
jobs.ashbyhq.com/periodic-labs/…
6. @PrimeIntellect — Head of Marketing ($150K–$300K)
jobs.ashbyhq.com/primeintellect…
7. @OpenAI — Content Studio Lead ($287K–$318K)
jobs.ashbyhq.com/openai/7ded0fa…
8. @AnthropicAI — Internal Communications Manager, Tech ($295K–$345K)
job-boards.greenhouse.io/anthropic/jobs…
9. @thinkymachines — Product Manager - Post Training ($350K–$400K)
jobs.ashbyhq.com/thinkingmachin…
10. @bfl_ai — Brand & Marketing Designer ($219K–$265K)
jobs.ashbyhq.com/black-forest-l…
11. @AnthropicAI — Executive Communications Writer ($295K–$345K)
job-boards.greenhouse.io/anthropic/jobs…
12. @thinkymachines — Global Public Policy ($300K–$360K)
jobs.ashbyhq.com/thinkingmachin…
13. @OpenAI — Product Designer, Youth ($245K–$310K)
jobs.ashbyhq.com/openai/43ba8b5…
14. @theworldlabs — Executive Business Partner (Office of the CEO) ($200K–$250K)
jobs.ashbyhq.com/worldlabs/28be…
15. @AnthropicAI — Head of Community, Enterprise Marketing ($400K)
job-boards.greenhouse.io/anthropic/jobs…
16. @SpaceXSI — Director of Global Growth Marketing and Sales Enablement ($200K–$280K)
job-boards.greenhouse.io/xai/jobs/52399…
17. @thinkymachines — Program Manager, Supercomputing ($350K–$475K)
jobs.ashbyhq.com/thinkingmachin…
18. @character_ai — Program Manager, Product Experience ($200K–$300K)
jobs.ashbyhq.com/character/56dc…
19. @OpenAI — Pricing Strategist, GTM ($234K–$260K)
jobs.ashbyhq.com/openai/5ce0931…
20. @AnthropicAI — Manager, Commercial Counsel, GTM ($320K–$385K)
job-boards.greenhouse.io/anthropic/jobs…
21. @periodiclabs — Controller ($250K–$300K)
jobs.ashbyhq.com/periodic-labs/…
22. @magicailabs — Founding Designer ($160K–$200K)
jobs.ashbyhq.com/magic/24c76152…
23. @PrimeIntellect — Head of People ($250K–$300K)
jobs.ashbyhq.com/primeintellect…
24. @periodiclabs — Business Operations, Product & Science ($250K–$300K)
jobs.ashbyhq.com/periodic-labs/…
25. @cohere — Executive Producer, Executive Communications ($150K–$210K)
jobs.ashbyhq.com/cohere/e645a8b…
0
0
4
247
Craig Hughes 𓅪 retweeted
I am working on a map of the secure program synthesis space, in collaboration with @sergeybratus for our NDIST keynote on Tuesday. Here's a sneak-peak. The words "RISC-V" and "Lean" are used here WLOG, and some simplifications are made to fit everything on one page. Did I miss emantics for each language in an interactive theorem prover, and then use these semantics to prove that every program is correct. (@oath_tech has done some work along these lines. See in particular Sam Xifaras' project on Ruby samx.io/blog/topics/de…)
2. Develop semantics for an architecture, say RISC-V, or perhaps for all possible architectures, and then use this to prove that every compiled program is correct. (@AnthropicAI made headlines recently with their announcement of a very large-scale project along these lines in collaboration with @theoremlabs anthropic.com/research/forma…)
3. Develop semantics for an architecture, say RISC-V, or perhaps for all possible architectures, and then use this to capture a formal model of each existing compiler or interpreter, and then iteratively (provably) refine each such model to an interpretable and easy-to-use formal semantics; then continue as in approach 1. (See foundational work by @kirancodes at BASIS AI basis.ai/blog/ship-your…)
4. Make Lean, or some other such ITP + programming language, fast to run (comparable to Rust or C), and then completely rewrite all software in it, and prove the rewrites are correct. (This is I think, loosely speaking, the idea behind @VictorTaelin 's Bend.)
5. Make implementation-level verification tools, such as Verus, more capable and ergonomic, and then use them to prove all software correct. (This is what @fhuszar and @Craig_Hughes_ are doing at Reasonable, & I'm aware that BASIS and AWS are also working on this.)
6. Design a language in which unsafe states are illegal, and then rewrite everything in that language. Here, what kinds of safety you can express will be much more limited but the problem is, conversely, way more tractable. (This is what @danwallach and @DARPA did with TRACTOR, where the safety property is memory safety.)
2. Develop semantics for an architecture, say RISC-V, or perhaps for all possible architectures, and then use this to prove that every compiled program is correct. (@AnthropicAI made headlines recently with their announcement of a very large-scale project along these lines in collaboration with @theoremlabs anthropic.com/research/forma…)
3. Develop semantics for an architecture, say RISC-V, or perhaps for all possible architectures, and then use this to capture a formal model of each existing compiler or interpreter, and then iteratively (provably) refine each such model to an interpretable and easy-to-use formal semantics; then continue as in approach 1. (See foundational work by @kirancodes at BASIS AI basis.ai/blog/ship-your…)
4. Make Lean, or some other such ITP + programming language, fast to run (comparable to Rust or C), and then completely rewrite all software in it, and prove the rewrites are correct. (This is I think, loosely speaking, the idea behind @VictorTaelin 's Bend.)
5. Make implementation-level verification tools, such as Verus, more capable and ergonomic, and then use them to prove all software correct. (This is what @fhuszar and @Craig_Hughes_ are doing at Reasonable, & I'm aware that BASIS and AWS are also working on this.)
6. Design a language in which unsafe states are illegal, and then rewrite everything in that language. Here, what kinds of safety you can express will be much more limited but the problem is, conversely, way more tractable. (This is what @danwallach and @DARPA did with TRACTOR, where the safety property is memory safety.)
6
23
129
8,375
‘the lanyard class’
@Craig_Hughes_ Posh and pauper should unite against the real enemy, the lanyard class.
0
0
1
416
this is "The internet's Wuhan moment"
@azeem on the week that marked a turning point
exponentialview.co/p/the-internet…
"The other half of the cure is proof."
@azeem on the week that marked a turning point
exponentialview.co/p/the-internet…
"The other half of the cure is proof."
0
0
6
317
Craig Hughes 𓅪 retweeted
The internet’s Wuhan moment
open.substack.com/pub/exponentia…
open.substack.com/pub/exponentia…
0
13
38
16,312
Software is dead, long live software!
What a week. GTC Berlin delivered. Exploits discovered. Formal verification promoted to essential status.
What a week. GTC Berlin delivered. Exploits discovered. Formal verification promoted to essential status.
1
1
6
403
Craig Hughes 𓅪 retweeted
Oh, you aren’t updating your dependencies every hour? Well, too bad, you are vulnerable
Oh, you are updating your dependencies every hour? Well, too bad, you are somehow also vulnerable
Oh, you are updating your dependencies every hour? Well, too bad, you are somehow also vulnerable
🚨 JFrog Security Research has identified a supply-chain attack involving malicious code in the LiteLLM 1.107.0 PyPI release. Avoid this version and investigate environments where it was installed or executed. More details will be shared soon in our blog post.
35
231
4,057
79,372
Craig Hughes 𓅪 retweeted
If you run a service on the internet and you aren't running a tool that mines your codebase for vulns right now you are committing malpractice.
Make sure you are using something built post-AI. Anything pattern-based / rule-based is not good enough. A quarterly pentest is not good enough.
You have a responsibility to your users. Don't be tomorrow's story.
Make sure you are using something built post-AI. Anything pattern-based / rule-based is not good enough. A quarterly pentest is not good enough.
You have a responsibility to your users. Don't be tomorrow's story.
8
4
46
5,324
At GTC in Berlin? Big boss says come visit us
see how formal verification is giving correctness guarantees for software
see how formal verification is giving correctness guarantees for software
1
3
6
859