Max Kerr @maxtalcai
CTO. Working on the dark art of synthetic data @ talc (YC S23). Formerly did privacy at Facebook. talc.ai Seattle, WA Joined July 2023-
Tweets188
-
Followers189
-
Following163
-
Likes461
this is definitely how a real non-propaganda business would behave finance.yahoo.com/news/exclusive…
Bringing this up again now that we’re talking about phi-3
Bringing this up again now that we’re talking about phi-3
My theory about the "delve" thing: it's because most scale AI contractors are based in kenya and nigeria, and formal english used in those countries heavily influences the data used in RLHF/fine tuning.
@mattshumer_ We’ll get longer ones out. Also, with the new tokenizer the context window should be a bit longerif you are comparing to Llama 2.
YC’s underrated superpower is the ability to consistently bring on group partners like Dave.
YC’s underrated superpower is the ability to consistently bring on group partners like Dave.
Smesneath @SmesneathMbR
0 Followers 220 Following
jk – e/acc @after_ephemera
508 Followers 329 Following building your grandkids' internet @mysten_labs | ex @meta ar/vr, privacy | holistic systems, radical moderation, accelerated future
MadgeHemingway @Elk99d1ksZXBGGt
4 Followers 431 Following
Felicia Dezalia @FelicDezal
55 Followers 5K Following
Smough @Smough450639
50 Followers 2K Following
Leshawez @leshawez21502
2 Followers 422 Following
Patty Banther @pat_banth
62 Followers 5K Following
AnnGeorge @810F0OJkz3AbC
4 Followers 391 Following
Advait Ruia @Advait_Ruia
2K Followers 1K Following Cofounder at SuperTokens: Open source user auth (YC S20)
Candida Pietsch @CandidaP42807
72 Followers 5K Following
IsabelPepys @aaNxpq3nE89sF4
5 Followers 355 Following
Rosio Wide @RosioW35713
38 Followers 5K Following
Sheauryn @SheaurynbYAOG
1 Followers 332 Following
Jacob Somer @jacob_somer_
695 Followers 4K Following AI Enthusiast & Software Engineer 💻 Building intelligent systems that make a difference.
Hermine Dearinger @hermi_dearin
68 Followers 5K Following
ÀI Enthusiast @wisdomvaultai
34 Followers 353 Following
Jeannette Annonio @JeannetteA69140
87 Followers 5K Following
Torrie Dechico @DechicoTor87086
82 Followers 5K Following
Merrie Hegg @HeggMerrie5383
70 Followers 5K Following
Molly Redmond @MollyRedmo18022
1 Followers 891 Following
Meta @MetaMeowMeow
93 Followers 2K Following
Kyle Senkbeil @kylesenkbeil21
227 Followers 1K Following CEO & Founder of Buy-Box AI | Early-stage startup | Father of four
Ashutosh Mehra @ashutoshmehra
2K Followers 6K Following Senior Principal Scientist at Adobe. Working on Acrobat AI Assistant, LLMs, and document ML.
Effie Osterhouse @EffieO14910
79 Followers 5K Following
Matt Ahmann @mattahmann
345 Followers 2K Following MSF Candidate @vanderbiltu | Space🚀, Tech🖥️, AI 🤖, Biotech 🧬, Sustainable Energy ☀️, Finance 💹, College Sports Fan 🏀🏈🏟️
Doreen Broden @doree_brod
30 Followers 5K Following
davidad 🎇 @davidad
13K Followers 7K Following Programme Director @ARIA_research | accelerate mathematical modelling with AI and categorical systems theory » build safe transformative AI » cancel heat death
Ellis Brown @_ellisbrown
363 Followers 940 Following CS PhD Student @NYU_Courant | Prev: @CarnegieMellon, https://t.co/EkTzprGc4i, @VanderbiltU
Michael J.J. Tiffany @kubla
2K Followers 2K Following Hacker: https://t.co/KdcpTomFWR Cofounder: https://t.co/oL9kikzNB2 (infosec unicorn+centaur) Cofounder: https://t.co/b4xRTQUfKY (personal data sovereignty) I have magnificent friends
kaushikpatnaik @kaushikpatnaik
906 Followers 3K Following Generative Human. Lives and breathes AI agents.Current: building @ai_minion. Prev: AR/VR Research (VisionPro) @ 🍎. Time Sinks: Reading, Coding, KPop and Anime.
N @nwang888
112 Followers 578 Following cs + econ @dukeu, co-director @hackduke, head ice cream scooper @pingrdev
Albert Young @AlbertYoung_1
7 Followers 745 Following
Art K @artk_
971 Followers 5K Following Anti-malware professional. Be excellent to each other. Opinions are personal not those of my employer.
Jeremy Harrison @hunterscannon
8 Followers 310 Following
Ma Sheen @MaSheenUprising
8 Followers 1K Following “The programme will take me a little while to run.” Fook glanced impatiently at his watch.
fvd26 @fvd26
32 Followers 217 Following
davidryan @hellodavidryan
4K Followers 935 Following Researching open source strategy in quantum computing for the OSRG. Former Head of Product at Quantum Brilliance, devrel at Red Hat & SaaS founder. Hi 👋
Michael Hartman @hartmams13
440 Followers 2K Following Personal account/views are mine alone; Former Exec Dir of the CO Dept of Revenue. Former Canopy Growth strategy. Recovering banker. Retweets = thought provoking
Advait Ruia @Advait_Ruia
2K Followers 1K Following Cofounder at SuperTokens: Open source user auth (YC S20)
Hugh Zhang @ICLR '24 @hughbzhang
2K Followers 566 Following open source ai @scale_AI. co-created @gradientpub.
Noam Brown @polynoamial
35K Followers 612 Following Researching reasoning @OpenAI | Co-created Libratus/Pluribus, the first superhuman no-limit poker AIs | Co-created CICERO | PhD from @SCSatCMU
The Visualizer @thevisualizerai
150 Followers 2 Following Empowering Learning and Creativity through AI
derek guy @dieworkwear
867K Followers 965 Following Menswear writer. Editor at @putthison. Creator of @RLGoesHard. Bylines at The New York Times, The Washington Post, The Financial Times, Esquire, and Mr. Porter
Chip Huyen @chipro
92K Followers 442 Following Data processing on GPUs @VoltronData Designing ML Systems: https://t.co/G81hL2dWmr @designmlsys #AI x #GPU
Omar Khattab @lateinteraction
11K Followers 2K Following CS PhD candidate @StanfordNLP. 2022 Apple Scholar in AI/ML. Author of ColBERT (https://t.co/2ZtgXoa1np), DSPy (https://t.co/BH7WmMKDXR), & various retrieval & LM systems.
Rich Harang @rharang
3K Followers 702 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 *`.
Molly White @molly0xFFF
117K Followers 2K Following crypto researcher & critic, software engineer, wikipedian • @web3isgreat creator • subscribe to my newsletter at https://t.co/WftJCrCfSY
Dylan Patel @dylan522p
39K Followers 685 Following SemiAnalysis Boutique AI & Semiconductor Research and Consulting DMs are open for consulting, quotes, or to talk shop
Mike Lewis @ml_perception
6K Followers 229 Following Llama3 pre-training lead. Partially to blame for things like the Cicero Diplomacy bot, BART, RoBERTa, kNN-LM, top-k sampling & Deal Or No Deal.
Philipp Schmid @_philschmid
16K Followers 657 Following Tech Lead and LLMs at @huggingface 👨🏻💻 🤗 AWS ML Hero 🦸🏻 | Cloud & ML enthusiast | 📍Nuremberg | 🇩🇪 https://t.co/l1ppq3q3hk
Sharan Narang @sharan0909
2K Followers 254 Following LLMs and AI Research (Llama 2 & 3 lead) @Meta | ex @Google (PaLM lead, T5), ex @Baidu (Deep Speech 2, Sparse Neural Networks), ex @Nvidia
Aston Zhang @astonzhangAZ
5K Followers 92 Following Research Scientist at the #llama team of Meta Generative AI, designing and training large language models. Opinions are my own.
Ahmad Al-Dahle @Ahmad_Al_Dahle
4K Followers 53 Following #Girldad of twins. Leading GenAI @ Meta (llama, imagine, meta ai and more)
oliver hill @_oliver_hill
380 Followers 491 Following building digital armor for your attention https://t.co/QwRpENaT8y
Jeremy Howard @jeremyphoward
222K Followers 5K Following 🇦🇺 Co-founder: @AnswerDotAI & @FastDotAI ; Hon Professor: @UQSchoolITEE ; Digital Fellow: @Stanford
Dan Siroker @dsiroker
33K Followers 3K Following Co-Founder & CEO of @LimitlessAI: a personalized AI powered by what you’ve seen, said, or heard. Formerly @RewindAI. Co-Founder of @Optimizely (sold for $300M).
Ellis Brown @_ellisbrown
363 Followers 940 Following CS PhD Student @NYU_Courant | Prev: @CarnegieMellon, https://t.co/EkTzprGc4i, @VanderbiltU
davidad 🎇 @davidad
13K Followers 7K Following Programme Director @ARIA_research | accelerate mathematical modelling with AI and categorical systems theory » build safe transformative AI » cancel heat death
Yi Tay @YiTayML
29K Followers 97 Following chief scientist / cofounder @RekaAILabs 🫠 past: research scientist @google brain 🤯 currently learning to be a dad 🍼
AI at Meta @AIatMeta
535K Followers 256 Following Together with the AI community, we are pushing the boundaries of what’s possible through open science to create a more connected world.
Greg Lindahl @glindahl
652 Followers 173 Following Renaissance man, and yes, I mean that in the rhetoric, music, and dance way. Currently @commoncrawl; formerly @EHTelescope, blekko, PathScale
Soumith Chintala @soumithchintala
187K Followers 888 Following Cofounded and lead @PyTorch at Meta. Also dabble in robotics at NYU. AI is delicious when it is accessible and open-source.
the tiny corp @__tinygrad__
33K Followers 61 Following We make tinygrad. Our mission is to commoditize the petaflop.
. @calliped
320 Followers 612 Following
Joe Mayo @JoeMayo
14K Followers 6K Following Building @generellem - #AI, #opensource, and #startups. Writing my own content.
Armand Joulin @armandjoulin
4K Followers 346 Following principal researcher, @googledeepmind. ex director of emea at fair @metaai. mostly work on open projects: fasttext, dino, llama, gemma.
Zeyuan Allen-Zhu @ZeyuanAllenZhu
8K Followers 274 Following physics of language models @ Meta / FAIR IOI - USACO - MCM - ACM/ICPC - Codejam Tsinghua - MIT - Princeton/IAS - MSR - FAIR
Jonathan Chavez @pirchavez
529 Followers 2K Following engineer, llm evals @datadoghq; ex @googleai, @atratopago (yc, accel); offensive security, ИТМО, ITESM
Lucas Beyer (bl16) @giffmana
57K Followers 448 Following Researcher (Google DeepMind/Brain in Zürich, ex-RWTH Aachen), Gamer, Hacker, Belgian. Mostly gave up trying mastodon as [email protected]
Eli Sennesh @EliSennesh
214 Followers 487 Following NHP electrophysiology @VanderbiltU. Predictive coding, probabilistic programming, affective science. Abolish the value function! It's all taxis navigation!
Hassan Hayat 🔥 @TheSeaMouse
5K Followers 4K Following Building the AI assistant for all @ https://t.co/D4gDyw97gu
priya joseph @ayirpelle
4K Followers 5K Following geek, entrepreneur, 'I strictly color outside the lines!', opinions r my own indeed. @ayirpelle , universal handle at this time
Emily Li @EmilyLiJiayao
395 Followers 632 Following Researcher @carnegiemellon | Founder @acadiaai | Prev research @modern_ai, ML @ evolution_devices, Founder @ arquestssern | Data-centric & Multimodal AI
oliver.mannion @tweeshan
320 Followers 2K Following Software, machine learning and people. Good at reading your tweets. Likes are bookmarks. 🥜🍄🥚
Databricks Mosaic Res.. @DbrxMosaicAI
30K Followers 115 Following We remove the barriers to state-of-the-art generative AI model development and make data + AI available to all.
Shreya Shankar @sh_reya
39K Followers 593 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.
Paul Calcraft @paul_cal
2K Followers 4K Following AI is good & bad, actually. Tweeting about AI/ML methods, software dev, research, tech and society, social impact. 20yrs in tech, 10 in ML/AI, PhD in comp sci
Yiğit Konur @yigitkonur
17K Followers 1K Following tweets daily about LLMs and claiming that #AI will take your job and ruin your morale 🤡 co-founder @thinkbuddyai & @wopehq - ex-gm @zeoagency
Jonathan Frankle @jefrankle
16K Followers 684 Following Chief Scientist, Neural Networks @Databricks via MosaicML. PhD @MIT_CSAIL. BS/MS @PrincetonCS. DC area native. Making AI efficient for everyone at @DbrxMosaicAI
JR Urbane Network @JRUrbaneNetwork
16K Followers 195 Following Trains are like dragons: massive, majestic, mystical and capable of quickly taking you on an exotic adventure.
Tom Kocmi @KocmiTom
606 Followers 178 Following Senior researcher at Microsoft Translator (he/him) | AI Evaluation (LLMs, MT, Multilingulity)@teortaxesTex 100%, gpt4 pre-training finished 2 yrs ago, they've been working on this for years. Also, anything they show was probably completed at least 6 months ago unlike Meta where Zuck probably unplugged training and chucked the checkpoint directly to huggingface. OAI are way ahead
A lot of what people presume is fixed about airline tickets is a social fiction that the airline empowers tens of thousands of people to waive if you ask nicely. (And there are other social fictional elements, like “a ticket means you definitely have a seat on the named plane.”)
Life pro tip: when you get to the airport early, especially if you have a connection, go up to every gate going to your destination and ask if you can board. Legit works ~40% of the time
It seems like a big limitation w/ AI agents rn isn't the AI, it's that the surface for software to interact w/ other software has gotten pretty small Software on a phone is isolated to itself, and most modern services don't have public APIs. Agents would need to physically tap?
@bindureddy As with Inflection, 'softie' will eventually acqui-hire OAI for a song .. proving also that YC thinking works well for Zero to Series (A/B ..) but its demo-or-die culture is not a replacement for brilliant strat-operators as CEOs (e.g. Zuck)
90% of the “evals are broken” guys never ran a single eval in their life.
This is really good work. But it also implies that any benchmark dataset that is openly published can never be reused
Data contamination is a huge problem for LLM evals right now. At Scale, we created a new test set for GSM8k *from scratch* to measure overfitting and found evidence that some models (most notably Mistral and Phi) do substantially worse on this new test set compared to GSM8k.
Data contamination is a huge problem for LLM evals right now. At Scale, we created a new test set for GSM8k *from scratch* to measure overfitting and found evidence that some models (most notably Mistral and Phi) do substantially worse on this new test set compared to GSM8k.
How to be as "smart" as Auto-Regressive LLMs: - memorize lots of problem statements together with recipes on how to solve them. - to solve a new problem, retrieve the recipe whose problem statement superficially matches the new problem. - apply the recipe blindly and declare…
There’s an art to distilling these to the absolute minimal necessary text. The human brain can’t comprehend how stupid these things are without practice.
There’s an art to distilling these to the absolute minimal necessary text. The human brain can’t comprehend how stupid these things are without practice.
Google firing excellent software engineers with no explanation, just as its valuation reaches the all-time high: If this doesn't explain how the software engineering market is changing dramatically, I don't know what does. It's not a pretty reality: but better to know about it.
It's more cost-effective to not launch in Europe than to implement GDPR compliant deletion protocols in vector databases and immutable files
Memory is now available to all ChatGPT Plus users. Using Memory is easy: just start a new chat and tell ChatGPT anything you’d like it to remember. Memory can be turned on or off in settings and is not currently available in Europe or Korea. Team, Enterprise, and GPTs to come.
Although I retweeted this because it's interesting, I wouldn't let it deter you from going to an Ivy League school if you can. At most these managers are looking to avoid a particular type of Ivy League student. If you're not that type, this won't affect you.
Now confirmed with data. Forbes surveyed managers with hiring authority. When asked whether more/less likely to hire vs 5 years ago: Ivy League: 7% more likely; 33% say less likely Public univs: 42% more likely; 5% less likely Selective privates: 37% more likely; 5% less likely
@paulg entry level jobs are decided by hiring manager. Mid to high by your network. The network value of Ivy League (with very little extra curation needed) is not controversial. You can 'simulate' it by selecting your buddies at a state school .. that takes work.hard reality
i understand why they have the delta airlines ceo and seattle's mayor, but how did AWS end up making the cut?
This morning the Department of Homeland Security announced the establishment of the Artificial Intelligence Safety and Security Board. The 22 inaugural members include Sam Altman, Dario Amodei, Jensen Huang, Satya Nadella, Sundar Pichai and many others.
Scatter plot with top-left good and yolo axes is the new radar plots where ours surrounds everything.
Good morning: @SnowflakeDB’s new 480B parameter #LLM is made of 128 experts! It’s bigger than #Grok and is now the largest *fully open source (Apache 2.0* LLM! 🧵👇 how does it compare to Llama 3, Mixtral, and GPT4?
Conference talks are a complete waste of time. Thousands of productive hours spent sitting and doing nothing. This can all be watched 2x speed on youtube.
@goth600 they're literally packaging bad evals into another eval and then renaming it as "enterprise intelligence". they're literally making CDOs with AI evals. you can't make this shit up lmao x.com/iliasmiraoui/s…
Not even questioning the model but when did training compute start mattering so much for end-users? Also - enterprise intelligence makes it sound like the model doesn't actually work well but we benchmark-hacked on these few to make it look good?
🤯 What sorcery is this? microsoft/Phi-3-mini-128k-instruct got the same scores in my LLM Comparison/Test as Llama 3 70B Instruct IQ1_S (16/18 + 13/18)! Llama 3 8B Instruct unquantized got 17/18 + 9/18! And phi-2-super and dolphin-2_6-phi-2 got 0/18 + 1/18 or 0/18 a while ago. 🤯
Trends for United States
You might like
