Adelin Travers @alkae_t
Principal Security Engineer, Machine Learning @ Trail of Bits, Views my own. alkaet.github.io Tokyo-to, Japan Joined June 2019-
Tweets208
-
Followers177
-
Following417
-
Likes2K
We have been named a Leader in the @forrester Wave. Check out their Q2 2024 report for cybersecurity consulting services to learn why @trailofbits scored among top vendors. trailofbits.com/forrester/
Interested in a postdoc position in Confidential Computing and Trusted ML on edge devices? Got a few days left to apply to us for a 3-year position as part of my @EPSRC Fellowship @ICComputing and @ImperialX_AI imperial.ac.uk/jobs/search-jo…
“We’ve created a way to reduce hallucinations,” the whole LLM problem space is that they are vibe machines, that is literally their personality. If you want to use them, use them for tasks you don’t need six nines on or bound how often you’re willing to be wrong
This is delivered by the ML Assurance team: @LucasBourtoule, @suhackerr and myself!
This is delivered by the ML Assurance team: @LucasBourtoule, @suhackerr and myself!
Infosec officers…. 🤣
The sound you hear is the sound of thousands pentesters crying who’d reported this as a finding all those years ago and ignored as nobody would do that…. arstechnica.com/security/2024/…
The vulnerability scan has 400+ findings. I exploit the domain with nothing on it. When is this madness going to change? LOL
See you there!
Going to RSA? Come to our AI event to meet fellow AI security nerds and peer into the future: lu.ma/cl7m4nbd
There appears to be a mismatch between publishing criteria in AI conferences and "what actually works". It is easy to publish new mathematical constructs (e.g. new models, new layers, new modules, new losses), but as Apple's MM1 paper concludes: 1. Encoder Lesson: Image…
supposedly above average ICML reviews this year...something, something...preference tuned language model outputs have a positive bias...
Fun commentary on LeftoverLocals; Thanks @ParkinsonFrost! Let's secure GPUs!
Fun commentary on LeftoverLocals; Thanks @ParkinsonFrost! Let's secure GPUs!
ML researchers working on scaling laws:
A pilot study by academics at the University of Oxford found some care providers had been using generative AI chatbots such as ChatGPT and Bard to create care plans for people receiving care. That presents a potential risk to patient confidentiality theguardian.com/technology/202…
It will be impossible for AI companies to fulfill their carbon commitments and satisfy their AI power needs without reactors. So, by 2026, we will be in the middle of a huge argument about nuclear power. U.S. SMR regulations currently look like this: x.com/andrewcurran_/…
It will be impossible for AI companies to fulfill their carbon commitments and satisfy their AI power needs without reactors. So, by 2026, we will be in the middle of a huge argument about nuclear power. U.S. SMR regulations currently look like this: x.com/andrewcurran_/…
I had to try this myself. @trailofbits was apparently founded by @DanielMiessler and Elijah Savage, not @dguido and @alexsotirov. It is known for having created the fastest open-source password cracker in the world, @shellphish.
I had to try this myself. @trailofbits was apparently founded by @DanielMiessler and Elijah Savage, not @dguido and @alexsotirov. It is known for having created the fastest open-source password cracker in the world, @shellphish. https://t.co/9krXDJGzmC
Thank you to Schmidt Sciences for supporting my group's efforts towards producing and enforcing technical specifications to curb the societal risks of ML models. These ideas were developed in collaboration with my students! More details: schmidtsciences.org/ai2050-early-c…
We went even further into the ML file formats rabbit hole and also made this repository linked in the blog post. It is intended to list and organize the many file formats used in ML. Feedback and contributions appreciated! github.com/trailofbits/ml…
I wrote a blog post about new features my team and I added to Fickling. Fun fact: Fickling can now differentiate and identify the various PyTorch file formats out there.
I wrote a blog post about new features my team and I added to Fickling. Fun fact: Fickling can now differentiate and identify the various PyTorch file formats out there.
Nalouse @nalouse49481
0 Followers 165 Following My hobbies are reading, food and sports. I like cats😘 I like to meet new friends while traveling🎉🎉🎉LLM Security @llm_sec
8K Followers 297 Following Research, papers, jobs, and news on large language model security. Got something relevant? DM / tag @llm_secNaman Devnani @naman_devnani
331 Followers 5K Following Security Researcher | Purple Team | Bug Hunter | CTF Player | Science & Tech Enthusiast | R&D | All-Source Intelligence | CAP | DCSP | TTIA | BCDEAleks Petrov @AleksPPetrov
168 Followers 447 Following PhD student at @OxfordTVG, @UniofOxford. I am working on Trustworthy Machine Learning, championing friendly AI and polite robotsEgan @7CH29fOt6C6UPe
198 Followers 1K Following 風雨を経験しなければ、虹を見ることはできない。磨かれなければ、ダイヤモンドになることはできない。堅持して、奮闘して、夢はついに実現します!頑張って!Kyle Easterly @KyleEasterly
123 Followers 1K Following Chief Technology Officer at Pango Technology • Founder @PromptLabAI / https://t.co/wvy47EYiiU • Software Architecture, Development, and Management • Generative AIEIIZ @elizi893
841 Followers 1K FollowingJoshua Saxe @joshua_saxe
3K Followers 982 Following AI+cybersecurity at Meta; past lives in academic history, labor / community organizing, classical/jazz piano, hacking scenec4lvin | ChainLight @SiwonHuh
616 Followers 1K Following Web3 Research Analyst @chainlight_io, @theori_io | c4lvin.ethak1010 @ak1010
1K Followers 987 FollowingDabby Dabs @Dabby_Dabs
766 Followers 5K Following You're on your own. And you know what you know. And you are the one who'll decide where to go. INTP 6'4" 👉🏽https://t.co/FRZab29W3H浜名佐代子 @eggqueen570787
8 Followers 78 FollowingTanya Davidson @TanyawyDavidson
31 Followers 924 Following Lured the best beauties from all US states 😜 Ready for a 1 on 1 meeting Some have naked profile photos! Watching this https://t.co/7YpXekoYQbWillow @teesmaez58237
120 Followers 1K FollowingTara @Thiffi287326
131 Followers 1K FollowingPetoust @petoust97725
29 Followers 706 Followingthrelfall @WHITEHACKSEC
415 Followers 341 Following working at intersection of offensive security, ai/ml & supply chains. enshittening the web @ https://t.co/zulqbxDZQV & https://t.co/EyMIpzuHUQchrisrohlf @chrisrohlf
11K Followers 783 Following 🇺🇸 Waging algorithmic warfare since 2003. Software and Security Engineer. Non-Resident Research Fellow @CSETGeorgetown CyberAITeemu Summanen @teemusum
195 Followers 3K Following Interested in AI, security, healthcare, and Flutter & Dart.👨🏼💻At X for reading diverse views by professionals and hobbyists.🔬📚🫶Shevad @Shevad512477
95 Followers 927 FollowingSparrow @Theau5095750270
293 Followers 5K Following See the world on the road, and get to know yourself on the way!Tyler Sorensen @Tyler_UCSC
982 Followers 842 Following Assistant Professor at UC Santa Cruz in CS and Security Research Engineer at @trailofbits Interested in PL and architecture design for heterogeneous systemsNasoot @nasoot79190
236 Followers 5K FollowingEvan Sultanik @ESultanik
1K Followers 477 Following Ph.D. computer security researcher @TrailOfBits. Editor of and frequent contributor to #pocorgtfo. My CV is a PDF that’s also an NES ROM https://t.co/lDrC4Hz6AIOsato Avan-Nomayo @OsatoNomayo
2K Followers 2K Following DeFi Correspondent @DLNewsInfo || Telegram: Osato || Scrabble and Poker hobbyistTrail of Bits @trailofbits
32K Followers 247 Following We help secure the world’s most targeted organizations and products. We combine security research with an attacker mentality to reduce risk and fortify code.Yessenia @corral24yesseni
219 Followers 3K FollowingRuth @ruth_joshua_
225 Followers 3K FollowingZea @Teesush174618
454 Followers 5K Following See the world on the road, and get to know yourself on the way!Michael Brown @MichaelBrownUC
273 Followers 789 Following Security Researcher, Premier @CincyLight fan accountIsmael Valenzuela @aboutsecurity
17K Followers 9K Following VP Threat Research & Intelligence @BlackBerry ▪️ Ex @Foundstone @Intel @McAfee ▪️ SANS Senior Instructor, GSE #132 ▪️ Author #SEC530 #SEC568 #ThinkRedActBlueanusha :D @0xk4l1
104 Followers 152 Following ai nerd. here for a good time. student @illinoiscs @sigpwny @aivillage_dcJoanna Miln @JoannagbMiln
18 Followers 489 Following Gathered on the site of girls from all US states 😍 Ready for private meetings Some have naked profile photos! Watching this https://t.co/8DiZEWtukrSheleath @Sheleath125456
25 Followers 324 Following "Your body is your most valuable asset. Treat it well, stay active, and enjoy the rewards of a fit and healthy life."Tothithe @tothithe79036
130 Followers 493 FollowingSoughez @Soughez177312
16 Followers 285 FollowingSlotosee @Slotosee110570
73 Followers 266 FollowingTachoana @Tachoana187219
26 Followers 308 FollowingTothyez @tothyez61129
38 Followers 292 FollowingMcNeeshe @McNeeshe1147
32 Followers 263 FollowingStoughsm @Stoughsm222831
29 Followers 288 FollowingTeslysh @teslysh54079
6 Followers 174 FollowingThesad @Thesad35188620
7 Followers 325 Following まだ時間があるうちに、自分が一番やりたいことを頑張って、なりたい自分になってください、世界はあなたが思っている以上に面白いものです。vicki @vboykis
52K Followers 1K Following Born: USSR. Raised: USA. ML Eng @mozillaai Ex: @duosec @Tumblr, @automattic Nights: 👦 & 👧 working on some ✨ new vectors ✨Halvar Flake @halvarflake
44K Followers 3K Following I do math. And was once asked by R. Morris Sr. : "For whom?" @[email protected] At the moment, for noone.Stella Biderman @BlancheMinerva
15K Followers 748 Following Open source LLMs and interpretability research at @BoozAllen and @AiEleuther. My employers disown my tweets. She/herAleks Petrov @AleksPPetrov
168 Followers 447 Following PhD student at @OxfordTVG, @UniofOxford. I am working on Trustworthy Machine Learning, championing friendly AI and polite robotsderek guy @dieworkwear
840K Followers 963 Following Menswear writer. Editor at @putthison. Creator of @RLGoesHard. Bylines at The New York Times, The Washington Post, The Financial Times, Esquire, and Mr. PorterJoshua Saxe @joshua_saxe
3K Followers 982 Following AI+cybersecurity at Meta; past lives in academic history, labor / community organizing, classical/jazz piano, hacking sceneNicole Perlroth @nicoleperlroth
92K Followers 6K Following cyber raconteur | author "This Is How They Tell Me The World Ends" | prev: digital espionage/sabotage @nytimes | now: putting all those words to workclem 🤗 @ClementDelangue
91K Followers 5K Following Co-founder & CEO @HuggingFace 🤗, the open and collaborative platform for AI buildersak1010 @ak1010
1K Followers 987 FollowingSakana AI @SakanaAILabs
19K Followers 0 Following We are a Tokyo-based R&D company on a quest to create a new kind of foundational AI model based on nature-inspired intelligence. https://t.co/LonvHEtlJRYanick Fratantonio (@.. @reyammer
7K Followers 749 Following Senior Research Scientist @ Google. Prev: Prof @EURECOM, Cisco Talos, DEFCON CTF Org. MOBISEC class: https://t.co/xPcWDxIcNo, Shellphish/NOPS hacker. Tweets mine.Netsec Explained @GTKlondike
1K Followers 475 Following I'm a senior security consultant who makes videos to level up my team on AI, pentesting, and bug bounty. Check out my channel on YouTube.threlfall @WHITEHACKSEC
415 Followers 341 Following working at intersection of offensive security, ai/ml & supply chains. enshittening the web @ https://t.co/zulqbxDZQV & https://t.co/EyMIpzuHUQHongyu (Charlie) Chen @hongyucharlie
76 Followers 283 Following share about personal learning and growthchrisrohlf @chrisrohlf
11K Followers 783 Following 🇺🇸 Waging algorithmic warfare since 2003. Software and Security Engineer. Non-Resident Research Fellow @CSETGeorgetown CyberAIJim Miller @TheMijCipher
343 Followers 151 Following Engineering Director for Cryptography at Trail of BitsXander Davies @alxndrdavies
1K Followers 478 Following technical staff @ uk ai safety institute prev student @harvard, director https://t.co/695XYMJSua, safety research with @davidbau and @DavidSKruegerTyler Sorensen @Tyler_UCSC
982 Followers 842 Following Assistant Professor at UC Santa Cruz in CS and Security Research Engineer at @trailofbits Interested in PL and architecture design for heterogeneous systemsMeredith Whittaker @mer__edith
92K Followers 4K Following President of @signalapp, Chief Advisor to @ainowinstitute (Also on Mastodon @[email protected], also on bsky @meredithmeredith.bsky.social)Chris Dahlheimer @tweet_c_d
59 Followers 158 Following Senior Sales Engineer at Trail of Bits @trailofbits. Tweets on #CyberSecurity and #SoftwareSecurity.Gene Meltser @gmeltser
183 Followers 195 Following Help you put cyber into your cyber, so you can cyber while you cyberJosselin Feist @Montyly
3K Followers 691 Following Engineering director at @trailofbits. Working on blockchain security & program analysisEvan Sultanik @ESultanik
1K Followers 477 Following Ph.D. computer security researcher @TrailOfBits. Editor of and frequent contributor to #pocorgtfo. My CV is a PDF that’s also an NES ROM https://t.co/lDrC4Hz6AIAccidental CISO @AccidentalCISO
54K Followers 2K Following I accidentally became the CISO. I didn't want this job, but the job chose me. I'm scared, and I want to go home.Aisha Alaagib @AishaAlaagib
668 Followers 1K Following ML researcher. Alumna: @AIMS_Next, @vectorinstitute, https://t.co/jBY6Y74H5o, @Mila_Quebec. ML communities: @DeepIndaba, @black_in_ai. Co-organizer @WiMLDSKhartoumdreadnode @dreadnode
783 Followers 22 Following AI Red Teaming | Research. Tooling. Evals. Cyber ranges.Michael Brown @MichaelBrownUC
273 Followers 789 Following Security Researcher, Premier @CincyLight fan accountTrail of Bits @trailofbits
32K Followers 247 Following We help secure the world’s most targeted organizations and products. We combine security research with an attacker mentality to reduce risk and fortify code.Ismael Valenzuela @aboutsecurity
17K Followers 9K Following VP Threat Research & Intelligence @BlackBerry ▪️ Ex @Foundstone @Intel @McAfee ▪️ SANS Senior Instructor, GSE #132 ▪️ Author #SEC530 #SEC568 #ThinkRedActBlueanusha :D @0xk4l1
104 Followers 152 Following ai nerd. here for a good time. student @illinoiscs @sigpwny @aivillage_dcAlexis Conneau @alex_conneau
24K Followers 113 Following Audio AGI Research Lead @OpenAI - GPT-Next - Past: XLM, Unsupervised ASR, Unsupervised MT, Wav2vec 2.0/XLSR, MUSE, Unsupervised cross-lingual transferRicky Skiro @RickySkiro
449 Followers 983 Following Security Engineer, Research @trailofbits, CTF player, passionate about security, member of CodeJitsu team @ CGC, part of the Core team of @rizinorgCarter Miller @CarterToB
132 Followers 223 Following Senior Technical Recruiter, Cyber Security for Trail of Bits! Passionate about all things cyber. Check out https://t.co/Kl92CHis8s!kache (dingboard.com) @yacineMTB
53K Followers 3K Following i'm a swe. go to https://t.co/pWRBfY8kn2 - AI image editing IN YOUR BROWSER! follow to watch a self funded founder beat VC backed AI startups with @dingboard_shane caldwell @shncldwll
352 Followers 259 Following pentester + ml eng. building hackbots @runsybilFredrik Dahlgren @fegge
132 Followers 108 Following All tweets are pre-approved by the Swedish monarchy. Also at @[email protected].ColdwaterQ (@coldwate.. @ColdwaterQ
122 Followers 75 Following Focused on Threat Research with an emphasis in AI and ML technologies. https://t.co/KfdoJc8vtlJoe Lucas @josephtlucas
412 Followers 1K FollowingmRr3b00t @UK_Daniel_Card
92K Followers 7K Following 真理的揭露者 Quis custodiet ipsos custodes fella in cyberspace #nafo undercover #FVEY Lovely Horse #fella #meme #farm #appreciator #cyber #specialistChris Nickerson @indi303
17K Followers 6K Following I build global cybersecurity companies, Red Teamer, CxO, Investor, Speaker, Mentor, Student, & Sprirt Animal . I'm also a liability #Breachquest #Damovo #LaresThomas Wolf @Thom_Wolf
68K Followers 4K Following Co-founder and CSO @HuggingFace - open-source and open-scienceAdam Dziedzic @adam_dziedzic
150 Followers 65 Following I'm a researcher, software developer, systems designer & engineer. I have a passion for machine/deep learning, databases, technology, traveling, sport & music.Joshua Wright @joswr1ght
28K Followers 555 Following Hacker for @counterhacksec and SANS Faculty Fellow. Pirata informático. Photography at https://t.co/Qbh3jsScLb. He/him.Rich Seymour @rseymour
2K Followers 3K Following working on a new thing past: @phosphorusinc, @CrowdStrike, @elastic, @endgameinc, hpc & matsci, opinions here solely mine, he/him 🦀 🎸François Chollet @fchollet
470K Followers 769 Following Deep learning @google. Creator of Keras. Author of 'Deep Learning with Python'. Opinions are my own.Nathan Hamiel @nathanhamiel
3K Followers 639 Following Sr Director of Research. BH Review Board Member and International Public Speaker. I focus on emerging technologies and future threats. @[email protected]Blue Team News @blueteamsec1
44K Followers 9K Following The cybersecurity home for the latest #BlueTeam, #DFIR, and #ThreatHunting news and tools.an0n @an0n_r0
11K Followers 716 Following CRT(E|O|L) | OSCP | @RingZer0_CTF 1st (for 2yrs) | HackTheBox Top10 | RPISEC MBE | Flare-On completer | GoogleCTF writeup winner | SSD research | Math MSc |🇭🇺We have been named a Leader in the @forrester Wave. Check out their Q2 2024 report for cybersecurity consulting services to learn why @trailofbits scored among top vendors. trailofbits.com/forrester/
@FuzzySec @haxorthematrix @yarden_shafir @AlanSguigna @jstrosch If you'd like to be a guest on the stream, please DM me on Twitter or on our Discord server!
Great streams coming up on the Off By One Security Channel from @FuzzySec, @haxorthematrix, @yarden_shafir, @AlanSguigna, @jstrosch, and more coming soon!! Please let me know what additional topics you'd like to see? Join our Discord Server here: discord.gg/offbyonesecuri…
Who's making the "i get my 0-days from arXiv" stickers for AI Village this year?
@alkae_t @suhackerr @moyix @trailofbits Maybe please join the OpenSSF working group and we can work on what solutions can be reused and what needs to be created from scratch. openssf.org/blog/2024/04/2…
@joshua_saxe Being a Security person, you may also enjoy “Demon of our own Design”. Makes a fairly basic point, but has kind of stuck with me. Have always wanted to work on “pricing risk” in the context of network. Lots of useful paradigms in finance. Maybe when I grow up.
@Cybrcrepidarian I don’t fully agree with what I wrote either - like everything in life - it depends 😆
@UK_Daniel_Card So often the difference between the smart intern and the seasoned pro isn't their understanding of topics, it's that gut feeling based on knowing what good and bad generally feel like, the ability to sniff out bad smells. You can't teach that, it comes with experience & exposure
people who worked in Helpdesk include: @GossiTheDog @SwiftOnSecurity and me :P (and fuck tons of other people) is this the only route? no. did I find it really helpful/useful? HELL YES.
Interested in a postdoc position in Confidential Computing and Trusted ML on edge devices? Got a few days left to apply to us for a 3-year position as part of my @EPSRC Fellowship @ICComputing and @ImperialX_AI imperial.ac.uk/jobs/search-jo…
“We’ve created a way to reduce hallucinations,” the whole LLM problem space is that they are vibe machines, that is literally their personality. If you want to use them, use them for tasks you don’t need six nines on or bound how often you’re willing to be wrong
@alkae_t @suhackerr @moyix @trailofbits That is true, neither signature, nor provenance are a silver bullet. Neither in ML, nor in traditional software. However, they do solve a class of risk and give you some foundation of trust.
@alkae_t @mihaimaruseac @moyix @trailofbits Yeah there’s a lot of model serialization formats. I started this repository to list and organize them (github.com/trailofbits/ml…). Some formats minimize the attack surface by not storing the architecture whatsoever or storing the weights separate from the architecture.
@mihaimaruseac @moyix @trailofbits They don’t (last I checked).
@mihaimaruseac @moyix @trailofbits conducted a security audit of safetensors that you can find here if you’d like some assurance: github.com/trailofbits/pu…. The file format itself is a header with JSON followed by tensors in binary.
At least stack overflow never snitched on us like this. Now there’s proof that I have no idea how to invert a binary tree.
GitHub Copilot Metrics API is now available in public beta! See your organization's aggregated usage metrics for Copilot completions and Copilot Chat 👇 github.blog/changelog/2024…
@joshua_saxe I think in practice it's a math-y way of saying "strictly bounded losses with unlimited potential gains are good" with the add-on (maybe from Black Swan?) that people are really bad at estimating low probabilities, so there's often arbitrage opportunity in those long-tail gains.