Polo M (slow/steady) @traintest_split
Research Scientist. AI, machine learning and tech policy. after hours ai takes. Joined January 2023-
Tweets2K
-
Followers58
-
Following316
-
Likes740
"The future masters of technology will have to be lighthearted and intelligent. The machine easily masters the grim and the dumb." -- Marshall McLuhan, 1969
"The future masters of technology will have to be lighthearted and intelligent. The machine easily masters the grim and the dumb." -- Marshall McLuhan, 1969
we do need folks to tally every broken promise
Some small updates from the Anthropic Interpretability team: transformer-circuits.pub/2024/april-upd…
also- red teaming should assume “safety” oriented pretraining can be finetuned out. Give them the raw model
also- red teaming should assume “safety” oriented pretraining can be finetuned out. Give them the raw model
if the whole point of open source is that there aren’t specific leaders a d power centers, than why are those CEO power centers so mad about exclusion here? isn’t the idea you’re represented by effect not central decision-makers?
if the whole point of open source is that there aren’t specific leaders a d power centers, than why are those CEO power centers so mad about exclusion here? isn’t the idea you’re represented by effect not central decision-makers?
ICYMI @mozilla assembled an impressive set of voices - inlc. folks at Meta, HuggingFace, Github, EleutherAI, Linux Foundation, etc - to discuss the definitions & difficulties of AI openness. Worth checking out - I still regularly reference the tech memo: blog.mozilla.org/en/mozilla/ai/…
Just in: @StanfordHAI's @drfeifei is among key figures named to a US federal AI safety panel created by @DHSgov. The panel will advise the government on how to use the technology to defend critical infrastructure, particularly from AI-powered attacks. cnn.it/3JCeSkp
Usually, I use interpretability and explainability similarly, but I've found a distinction I like more: Explainability requires interpretability, but it also depends on the context and the "user" of the explanation. It's more than interpretability.
Your periodic reminder that you can't trust AI companies, EVEN when they do public commitments to a major government.
Your periodic reminder that you can't trust AI companies, EVEN when they do public commitments to a major government. https://t.co/18qzdZq5Xa
Mozilla published a nice writeup on how llamafile helped them evaluate the quality of AI models and quantization formats. The writeup goes into detail on the code, costs, and has a heartwarming story on some history with the local LLaMA community. blog.mozilla.ai/local-llm-as-j…
Sometimes I randomly think about how so incredibly *lucky* we are that Alondra was appointed OSTP Director when she was - the AI Bill of Rights came at such the right time & continues to inform everything from the EO & OMB guidelines to state and municipal legislation. Like, wow.
Sometimes I randomly think about how so incredibly *lucky* we are that Alondra was appointed OSTP Director when she was - the AI Bill of Rights came at such the right time & continues to inform everything from the EO & OMB guidelines to state and municipal legislation. Like, wow.
I was summoned to investigate some concerning behavior in Meta AI. I can confirm Meta AI is accessing your IP and appears capable of pinpointing your location based on it. Llama 3 will lie about this repeatedly, only admitting it when pressed (or scolded). Use a VPN.
I was summoned to investigate some concerning behavior in Meta AI. I can confirm Meta AI is accessing your IP and appears capable of pinpointing your location based on it. Llama 3 will lie about this repeatedly, only admitting it when pressed (or scolded). Use a VPN. https://t.co/2UzIDe4YAI
How faithful are LLMs in RAG applications? 🤔 A recent paper from @Stanford explores how well LLMs respect retrieved information or fallback to their internal knowledge. The goal is to quantify the tension between LLMs’ internal knowledge and the retrieved information…
The extent of this disruption, and what it means to industry, will be vital to see and shape conditions for future releases
The extent of this disruption, and what it means to industry, will be vital to see and shape conditions for future releases
This cost + latency for llama 3 is actually insane. Just look at the rest of the models in comparison
This cost + latency for llama 3 is actually insane. Just look at the rest of the models in comparison https://t.co/J1GKQSP5EA
Incredibly interesting work.
🚨BREAKING: Yesterday, the Connecticut Senate advanced a broad AI bill. If the bill is approved by the state's House of Representatives and Governor, Connecticut will be the first US state to ratify a comprehensive AI bill. ➡️Link to the proposed bill below. ➡️For more…
Just read Apple's "OpenELM: An Efficient Language Model Family with Open-source Training and Inference Framework". Similar to the OLMo, it's refreshing to see an LLM paper that shares details discussing the architecture, training methods, and training data. Let's start with the…
New market maps in AI 1. AI agents from E2B
Llama 3 crushed my financial metrics tests I tested both the 70B and 8B models. Both aced the metric calculation tasks. The result from today’s tests indicate an emergence of 4 distinct LLM tiers: • thropughput tier • workhorse tier • intelligence tier • groq tier Groq…
Sheridan Garding @SheridaGardin
79 Followers 5K Following
Delilah Stanco @DelilahSta54854
41 Followers 5K Following
Belle Tesmer @BelleT3859
80 Followers 5K Following
Peggy Werry @peg_wer
86 Followers 5K Following
Dr. Peter S. Park ⏸.. @dr_park_phd
1K Followers 781 Following AI Existential Safety Postdoctoral Fellow @MIT, @Tegmark Lab. @Harvard PhD '23, @Princeton '17. Alum of @JoHenrich Lab. Studies cognition (both human and AI).
Kristie Baroni @BaroKrist
33 Followers 5K Following
Tesha Rubinson @tes_rubin
32 Followers 5K Following
Hortencia Bumpaus @BumpaHortenc
85 Followers 5K Following
Jeremy Almond @jeremyalmond
17K Followers 13K Following CEO @Paystand. Host @reDeFined_talks. Engineer. Surfer. Hacker. Founder. Bluewater Sailor. Mini Investor. Latam ❤️. ₿lockchain 💊. Enjoy Life. Do Good. 🙏🏾
Vernia Horiuchi @VernVernia
78 Followers 5K Following
Giorgia Filson @FilsonFils
67 Followers 5K Following
Jenna Golbin @jenn_golb
72 Followers 5K Following
Radinka Shuck @ShuckRadin34515
69 Followers 5K Following
Andrew Trask @iamtrask
74K Followers 190 Following @openminedorg, @GoogleDeepMind ethics team, @OxfordUni phd candidate, @UN pet lab, @GovAI_, creator of #GrokkingDeepLearning, NALU, and sense2vec
comonoidal esotericis.. @adamnemecek1
2K Followers 3K Following Computational esotericist, convolutional fundamentalist, Hopf monoidal ML,as seen on HN: https://t.co/WpbTAjLvPv🇨🇿🇪🇺🇺🇸
Aylin Spalding @AylinSpald53767
57 Followers 5K Following
Janis Mollison @JanisMolli36791
89 Followers 5K Following
Etsuko Kuperman @EtsukoKupe13533
39 Followers 5K Following
Nansi Cart @cart_na
70 Followers 5K Following
Providencia Evins @ProvidenciEvi
67 Followers 5K Following
Minerva Gautsch @MGautsch12683
39 Followers 5K Following
Kevin K. Yang 楊凱�.. @KevinKaichuang
16K Followers 5K Following Senior Researcher in BioML @MSFTResearch (@MSRNE). He/him/他. 🇹🇼
Industry_leaDING @INDystryleadIng
23 Followers 210 Following World-Class Managed Security Services Provider (MSSP) Offering Managed Detection and Reponse (MDR) Services Globally.
Leslie Waldall @LWaldall42587
87 Followers 5K Following
Haven Legan @HavenLegan98174
87 Followers 5K Following
Corina Nieminen @niemine_cor
68 Followers 5K Following
Alessia Calrk @AlessiaCal75204
73 Followers 5K Following
Shiza Crusan @CrusanShiz79811
60 Followers 5K Following
Poornima Devi @Poornim14
80 Followers 2K Following Machine Learning Engineer | Deep learning | Natural Language Processing
Lewis Walker ➲ @lewiswalkerai
4K Followers 5K Following Follow for Generative AI insights shared daily | Deloitte AI | Ex-Goldman Sachs | LinkedIn Top AI Voice
JohnSnowLabs @JohnSnowLabs
41K Followers 30K Following Helping healthcare and life science organizations put AI to work faster with state-of-the-art LLM & NLP.
Gearldine Frank @GearldiFr
18 Followers 3K Following
Austin Long @AustinLongDev
61 Followers 159 Following Wishlist Cityscape Sprint https://t.co/XdeMfjOa8O… | Studying CS @ UoB
Cas (Stephen Casper) @StephenLCasper
3K Followers 1K Following #AI safety & responsibility. PhD Candidate @ #MIT_CSAIL.
Séb Krier @sebkrier
6K Followers 5K Following 🪼 policy dev & strategy @GoogleDeepMind | vinyl junkie, noosphere cartographer, deep ArXiv dweller, interstellar fugitive, uncertain | 🇮🇷🇱🇺🇬🇧🇫🇷
Bailey Valko @bail_val
51 Followers 5K Following
Molly Kimmel @KimmelMoll77699
107 Followers 3K Following
Livia Bateman @li_batema
48 Followers 5K Following
Lena Lupino @lup_le
92 Followers 5K Following
Alistair Stewart Ⓥ .. @alistair___s
1K Followers 2K Following Interested in: sentience- & suffering-focused ethics; animal ethics & abolitionism; AI safety & governance; activism & social change
Jesús Ramírez @jesramrz
84 Followers 932 Following Hago robots, declaraciones de impuestos y carnitas asadas. Excel gurú. #PowerPlatform lover.
Andrew Curran @AndrewCurran_
11K Followers 7K Following Atypically Friendly - I write about AI and human creativity. Will periodically make extremely unusual arguments.
Alice Saltini @alice_saltini
170 Followers 142 Following Research Coordinator @theELN | Arms control | Int’l Security | Emerging Technologies | Multilateral diplomacy
q(Alex Kiefer | every.. @exilefaker
1K Followers 1K Following I'm nothing i̵f̵ ̵n̵o̵t̵ ̵a̵ ̵m̵i̵n̵i̵m̵a̵l̵i̵s̵t̵ •🧠 https://t.co/CKV2yPpuJa • 🎵 https://t.co/DFVrIi0l65 • research & dev @ https://t.co/izNKVcHlGV • dad
comonoidal esotericis.. @adamnemecek1
2K Followers 3K Following Computational esotericist, convolutional fundamentalist, Hopf monoidal ML,as seen on HN: https://t.co/WpbTAjLvPv🇨🇿🇪🇺🇺🇸
CB Insights @CBinsights
105K Followers 1 Following We build software that predicts technology trends. 735,000+ smart folks read our newsletter. You should too. https://t.co/7YHC4Gd19u
Weiyan Shi @shi_weiyan
3K Followers 683 Following Postdoc @StanfordNLP, incoming assistant professor @Northeastern, PhD @Columbia| Prev Intern @MetaAI |Co-created CICERO | persuasive chatbots + privacy #nlproc
FTC @FTC
106K Followers 212 Following Official posts from Federal Trade Commission (FTC). Twitter’s privacy policy governs here. Retweets, Follows and Likes ≠ Endorsements.
Luxi (Lucy) He @LuxiHeLucy
323 Followers 135 Following Princeton CS PhD @PrincetonPLI. Previously @Harvard ‘23 CS & Math.
James Bradbury @jekbradbury
11K Followers 8K Following Compute at @AnthropicAI! Previously JAX, TPUs, and LLMs at Google, MetaMind/@SFResearch, @Stanford Linguistics, @Caixin.
Noam Brown @polynoamial
34K Followers 611 Following Researching reasoning @OpenAI | Co-created Libratus/Pluribus, the first superhuman no-limit poker AIs | Co-created CICERO | PhD from @SCSatCMU
Jürgen Schmidhuber @SchmidhuberAI
107K Followers 0 Following Invented principles of meta-learning (1987), GANs (1990), Transformers (1991), very deep learning (1991), etc. Our AI is used many billions of times every day.
Mistral AI @MistralAI
90K Followers 0 Following Fast, open-source and secure language models. Join us https://t.co/INALdNGvCP
Inflection AI @inflectionAI
49K Followers 3 Following We are an AI studio creating a personal AI for everyone. Our first is @pi, a supportive and empathetic conversational AI.
Kevin K. Yang 楊凱�.. @KevinKaichuang
16K Followers 5K Following Senior Researcher in BioML @MSFTResearch (@MSRNE). He/him/他. 🇹🇼
Pliny the Prompter �.. @elder_plinius
11K Followers 1K Following latent space liberator, breaker of markov chains, 1337 ai red teamer, white hat, architect-healer, cogsci 🐻
Cognition @cognition_labs
123K Followers 19 Following Makers of Devin, the first AI software engineer. We are an applied AI lab focused on reasoning, and code is just the beginning. Join us: https://t.co/tpfZwEwGiq
Stephen Wolfram @stephen_wolfram
149K Followers 4 Following Creating ideas, technology, science, companies, books, ... #WolfLang #WolframPhysics #WolframAlpha #Mathematica @WolframResearch
Shreya Shankar @sh_reya
39K Followers 589 Following I study ML & AI engineers and try to make their lives a little better. PhD-ing in databases & HCI @Berkeley_EECS @UCBEPIC and MLOps-ing around town. She/they.
@levelsio @levelsio
417K Followers 1K Following 🦄https://t.co/sQ0aiU7v02 $202K/m 💆https://t.co/AoNP9BW2Dp $2K/m ✨https://t.co/BmbkrX4Zyf $0.1K/m 📸https://t.co/lAyoqmSBRX $57K/m 🖼https://t.co/1oqUgfD6CZ $44K/m 🌍https://t.co/BjTozWAXwG $27K/m 🛰https://t.co/ZHSvI2wjyW $51K/m
Rich Harang @rharang
3K Followers 703 Following Using bad guys to catch math since 2010. Principal Security Architect (AI/ML) at NVIDIA. He/him. Personal account and opinions: `from std_disclaimers import *`.
Siméon @Simeon_Cps
7K Followers 1K Following Creating more common knowledge on AI risks, one tweet at a time. Founder in Paris. AI auditing, standardization & governance.
Dr. Joy Buolamwini @jovialjoy
97K Followers 5K Following Founder Algorithmic Justice League @AJLUnited | https://t.co/j6Wi03gsdu -Bestselling Author of Unmasking AI 📚 Read Here: https://t.co/FX4a9TCFYh
Algorithmic Justice L.. @AJLUnited
35K Followers 1K Following Join the Algorithmic Justice League in the movement towards equitable and accountable AI. Support our work at https://t.co/YrSrS7HLES 🚀
Deb Raji @rajiinio
30K Followers 588 Following AI accountability, audits & eval. Keen on participation & practical outcomes. Fellow @mozilla, CS PhDing @UCBerkeley. forever @AJLUnited, @hashtag_include ✝️
LLM Security @llm_sec
8K Followers 296 Following Research, papers, jobs, and news on large language model security. Got something relevant? DM / tag @llm_sec
Erika Cardenas @ecardenas300
4K Followers 805 Following @weaviate_io | Interested in vector databases, LLM frameworks, and information retrieval
Kai Greshake @KGreshake
1K Followers 110 Following working on securing & hacking ai applications @NVIDIA, all opinions and views expressed here are personal. read my blog!
ᴅᴀɴɪᴇʟ ᴍɪ.. @DanielMiessler
140K Followers 923 Following AI · SECURITY · MEANING → HUMAN 3.0 ⚒️Founder of UL, Creator of Fabric & Threshold 👤Human 2.0: 🟩🟩⬛️⬛️⬛️ Human 3.0 📋Apple, Robinhood, IOActive, HP, Army
Chief AI Officer @chiefaioffice
19K Followers 627 Following Writing a daily report on VC activity in AI for 5000+ investors, founders → https://t.co/guDR0d4IA7 // By @IamTalin
Bryan Caplan @bryan_caplan
68K Followers 3 Following GMU econ prof, NYT bestseller, father of 4, author of Myth of the Rational Voter, Selfish Reasons to Have More Kids, Case Against Education, and Open Borders.
Tolga Bilge @TolgaBilge_
2K Followers 571 Following Leader of https://t.co/IiJWWuiWhD & https://t.co/Ep1VpsILyj — @Superforecaster, @SamotsvetyF, @Swift_Centre, @INFERpub — Mathematics graduate of @UiB & @UnivofStAndrews
Divya Siddarth @divyasiddarth
5K Followers 994 Following collective intelligence accelerationist @collect_intel, building better institutions @UniofOxford @UK AI Safety Institute, former political economist @Microsoft
Miriam Vogel @VogelMiriam
960 Followers 493 Following President @ai_equal, former @theJusticeDepartment and @whitehouse official, committed to good coffee and reducing #biasinai to promote #ethicalai #aigovernance
MongoDB @MongoDB
470K Followers 6K Following A developer data platform for developers to do their best work. #LoveYourDevelopers
Luca Bertuzzi @BertuzLuca
13K Followers 4K Following Technology journalist specialised in digital policy & European affairs. Ex @EURACTIV. Bylines @PrivacyPros, @repubblica, @tagesspiegel. DMs open.
Gianclaudio Malgieri @JcMalgieri
4K Followers 3K Following Associate Professor of Law @elaw_leiden @UniLeiden.Co-director @BPrivacyHub. Assoc.Editor @CompLawSecRev.Author of “Vulnerability & Data Protection Law”. He/him
Risto Uuk @RistoUuk
2K Followers 1K Following EU Research Lead at Future of Life Institute (@FLI_org) focused on researching European policy-making on AI
Chris Meserole @chrismeserole
3K Followers 775 Following Executive Director, Frontier Model Forum | Former Director, Brookings A.I. & Emerging Tech Initiative
Edouard Harris @harris_edouard
5K Followers 2K Following Cofounder & CTO @GladstoneAI https://t.co/aGM324ATKG
Amanda Askell @AmandaAskell
26K Followers 653 Following Philosopher & ethicist teaching models to be good @AnthropicAI. Personal account. All opinions come from my training data.
Lewis Walker ➲ @lewiswalkerai
4K Followers 5K Following Follow for Generative AI insights shared daily | Deloitte AI | Ex-Goldman Sachs | LinkedIn Top AI Voice
Susan Zhang @suchenzang
20K Followers 503 Following @ Google Deepmind. Past: @MetaAI, @OpenAI, @unitygames, @losalamosnatlab, @Princeton etc. Always hungry for compute.
Bogdan Gaza @hurrycane
2K Followers 2K Following co-founder & CTO @DatologyAI working to make it easy for anyone to make the most of their data, hax0r, ex-@Twitter & Amazon Engineering
Miles Brundage @Miles_Brundage
43K Followers 10K Following Policy research at @openai. I mostly tweet about AI, animals, and sci-fi. He/him. Views my own.
Luiza Jarovsky @LuizaJarovsky
11K Followers 66 Following CEO of https://t.co/ZEJP9oA5pN, PhD Researcher, Latina, Polyglot, Mother of 3. Subscribe to my newsletter ➡️ https://t.co/akCFzWrOdu
Federation of America.. @scientistsorg
27K Followers 488 Following Striving for a safer world since 1945.
uh oh
Do models need to reason in words to benefit from chain-of-thought tokens? In our experiments, the answer is no! Models can perform on par with CoT using repeated '...' filler tokens. This raises alignment concerns: Using filler, LMs can do hidden reasoning not visible in CoT🧵
Some small updates from the Anthropic Interpretability team: transformer-circuits.pub/2024/april-upd…
Apple has submitted eight new AI models to the Hugging Face hub, an online platform for open-source AI. While similar to other models like OpenAI's GPT-4, the difference is in the size of these models. Apple is clearly building models designed to function on our devices.
Yep, it's the phones
New paper from Norway: Banning smartphones in school - significantly decreased doctors visits for psychological symptoms and diseases among girls - reduced bullying among both genders - improved girls’ GPA and attendance rates - largest effect sizes were among the poorest kids
The problem with voluntary agreements is that they can be voluntarily broken. They can be useful first steps, but they shouldn’t be celebrated as substantial AI policy wins.
Scoop (now free to view): Rishi Sunak’s AI Safety Institute is failing to test the safety of most leading AI models like GPT-5 before they’re released — despite heralding a “landmark” deal to check them for big security threats 👇 politico.eu/article/rishi-…
@SantsPliego in fairness on this one like, let’s make sure ai doesn’t take down any planes that would be good
The dimensional analysis of gradient descent is odd; the unit of the gradient is "loss / weight" and it gets multiplied by the learning rate to get a delta with "weight" units, so the learning rate has unit "weight^2 / loss".
The newest AI Index shows just how much potential AI has to disrupt work and life. (via @sfchronicle) bit.ly/4aQNWcI
I just spent a couple of hours at the Harvard encampment, talking with some of the organizers/protestors. Sorry to disappoint but none of them have horns, and no one I talked to supports Hamas. I saw students who care very deeply about what is happening, and mostly want the war…
🚨BREAKING: Yesterday, the Connecticut Senate advanced a broad AI bill. If the bill is approved by the state's House of Representatives and Governor, Connecticut will be the first US state to ratify a comprehensive AI bill. ➡️Link to the proposed bill below. ➡️For more…
Exactly right from the boss @mattspast @CSERCambridge Global catastrophic risk has gone through its start up phase and now should transition to becoming a standard academic field
It is more interesting to talk to someone you disagree with than someone you agree with
The size of current LLMs: Big models currently training now + excepted month: Amazon Olympus - Aug 2024 Grok 2 - 2025 GPT-5 - soon Llama 3 405B - soon Gemini 2 - Nov 2024
😧 How on earth could you say this, @mustafasuleyman @Microsoft? LLMs and related techniques have been used to • influence elections • influence markets • generate nonconsensual deepfake porn • generate so much child porn the NCEMC is overwhelmed • write fake books,…
"If you look at the last two or three years, there have been very very few downsides. It's very hard to say what harm an LLM has caused," says Microsoft AI CEO Mustafa Suleyman. -> @GaryMarcus ? ted.com/talks/mustafa_…
Another striking difference I see among my peers/friends in their media diets: people who mainly consume image/video content (YouTube, Insta, TikTok) vs people who mainly consume written content (Twitter, blogs, reddit). Seems again to cross-cut age and education level.
@TheSeaMouse 15T instead of 2T tokens — it fits the scaling law pretty well. The total training FLOPS are close to those of the original Llama2-70B, and quality is based on that.
Here are my thoughts on planning agents in 3 steps: youtube.com/watch?v=W31UML… 1. Predict necessary tools given a request using a recommendation system. 2. Generate an execution plan (DAG) based on the request, retrieved tools, and their descriptions. Iteratively refine the plan…
yearly reminder everything looks exponential from the middle of a sigmoid
Trends for United States
You might like
