Principal Engineer at @allen_ai. Engineering Lead of the OLMo project.
Seattle, WA
Joined August 2018
- Tweets1.3K
- Following296
- Followers1K
- Likes446
Mechanical Dirk retweeted
We now have state-of-the-art pre-trained base models in the 3T-parameter weight class, built entirely in-house at MAI.
Post-training results coming soon -- stay tuned!
Lots of details in our technical blog post.
microsoft.ai/news/frontier-p…
I love the idea that there is someone at Anthropic whose job it is to set the "tone" of Claude, and when they see the feedback they go "Claude speak doesn't resonate with everyone — and honestly? That's valid."
I think it was @shanearora0 who first gerrymandered the tokens? I hope that becomes a term of art 🤣
We’re releasing Olmo-core 3—open training infrastructure for large mixture-of-experts (MoE) models.
It’s a core system behind the next generation of Olmo, designed to scale into the trillion-parameter range. 🧵
💻 GitHub repo: github.com/allenai/olmo-core
Mechanical Dirk retweeted
Today we're unveiling Trillium Labs @trillium_labs, a new non-profit to foster the open science of frontier AI. We're building open post-training recipes and will expand into open infra to study RSI, reward-hacking, multi-agent systems, and whatever comes next.
We're built around the theory of change that you need more eyes to solve hard technical problems. We have faith in the scientific methods and communities that humanity has built, and worry that AI is becoming too closed to utilize them.
Trilliums are wildflowers that bloom briefly in the spring, before the forest canopies fill out. Though they are small, they lay the foundation for the cycles of growth and nourishment through the rest of the year. At Trillium Labs, the recipes will be the slow nutrients for the seasons and the model releases will be the blooms. Building an institution dedicated to this is needed because, much as nature’s trilliums are slow to expand and grow, the open-ecosystem needs time and dedicated resources to catch up.
I co-founded with with a long-time friend and collaborator Tom Zick (@thesezickbeats). We're hiring (full time + student collabs/interns), we're fundraising, and we're looking for compute. Please get in touch if you're interested in helping out. Offices based in the Bay Area and Cambridge MA, remote okay.
I’m in the Bay Area until for The Curve and COLM to connect with people who are interested. We’re thankful to have initial support from Halcyon Futures and Schmidt Sciences with more funding en route to enable our ambitions of scaling. Our advisors @Thom_Wolf, @HannaHajishirzi, @gneubig and @ctnzr have been instrumental to building the ecosystem that exists today, and I’m stoked to get to keep working with them.
Mechanical Dirk retweeted
We’re releasing Olmo-core 3—open training infrastructure for large mixture-of-experts (MoE) models.
It’s a core system behind the next generation of Olmo, designed to scale into the trillion-parameter range. 🧵
💻 GitHub repo: github.com/allenai/olmo-core
Mechanical Dirk retweeted
>be me
>discover effective altruism
>apparently normal charity is inefficient
>why donate to random sad thing when spreadsheet can tell you optimal sad thing
>fair enough
>buy mosquito nets
>save lives
>numbers look good
>feel powerful
>couple years later
>someone asks an innocent question
>why only count people alive today
>huh
>future people matter too
>obviously
>my grandchildren shouldn't matter less just because they haven't spawned yet
>reasonable.jpg
>keep following logic
>what about their grandchildren
>also yes
>what about people in 500 years
>sure
>5000 years
>why not
>500 million years
>starting to get weird but morality is morality
>open calculator
>humanity could survive for an astronomically long time
>could colonize galaxy
>could have trillions upon trillions of descendants
>maybe digital people too
>maybe simulated civilizations
>maybe dyson spheres full of happy uploaded minds
>calculator starts smoking
>realize currently living humans are rounding error
>8 billion people suddenly looking extremely beta
>future contains potentially 10^something people
>can't even fit beneficiaries in google sheets
>new moral priority unlocked
>protect the long-term future
>stop thinking in units of "people helped"
>start thinking in "fraction of cosmic endowment preserved"
>malaria?
>terrible
>but only kills existing humans
>AI extinction could delete the entire light cone
>nuclear war could permanently derail civilization
>bad institutions could lock in terrible values for ten million years
>someone invents wrong constitution in 2140
>quadrillions suffer
>better fund governance workshop now
>friend says maybe we should improve hospitals
>explain opportunity cost
>friend says hospitals are full of actual sick people
>explain scope sensitivity
>friend stops inviting me to dinner
>need to decide what to fund
>easy
>expected value
>suppose project has one in a million chance of preventing extinction
>sounds tiny
>but extinction destroys 10^50 future lives
>multiply
>mother of god
>$10 million project has expected value of several galaxies
>charity evaluation complete
>someone asks where the one-in-a-million number came from
>expert judgement
>which expert
>us
>how calibrated
>extremely thoughtfully
>reduce estimate to one in ten million to be conservative
>still beats curing cancer by 38 orders of magnitude
>epistemic robustness achieved
>someone says maybe project doesn't work
>assign 20% chance
>still astronomical
>maybe project makes problem worse
>assign 5% chance
>still astronomical
>why 5
>because 30 felt pessimistic
>publish 46-page report
>contains seventeen sensitivity analyses
>every sensitivity analysis begins after assuming intervention has positive sign
>critic says you're multiplying enormous hypothetical stakes by extremely uncertain probabilities
>yes
>that's literally why it's important
>critic says the uncertainty might be structural rather than numerical
>make probability smaller
>critic says no, I mean maybe your model is wrong
>make probability smaller again
>critic begins rubbing temples
>discover AI safety
>perfect longtermist cause
>AI might kill everyone
>or create utopia
>or seize galaxy
>or tile universe with paperclips
>or create billions of conscious software minds
>finally a problem with numbers big enough for me
>start AI safety nonprofit
>mission: prevent dangerous AI
>hire smartest people available
>smartest people immediately start building better AI to understand dangerous AI
>interesting
>we must understand capabilities to understand safety
>we must scale models to study alignment
>we must race ahead so less responsible actors don't get there first
>we must deploy systems to learn how deployment can go wrong
>we must build the thing quickly because building the thing quickly is dangerous
>outsider asks why the people most worried about AI apocalypse all work at AI companies
>complicated field
>company releases stronger model
>very concerned
>company begins training even stronger model
>extremely concerned
>company raises $14 billion
>concern reaches unprecedented levels
>need to influence government
>future is at stake
>normal democratic process too slow
>politicians don't understand exponential curves
>public doesn't understand x-risk
>experts must guide them
>who counts as expert
>people who understand x-risk
>who understands x-risk
>our friends
>someone objects that this seems politically convenient
>explain we're representing future generations
>future generations unavailable for comment
>develop concept of value lock-in
>terrifying possibility that one ideology controls civilization forever
>therefore extremely important that civilization adopts correct values before lock-in
>whose values
>let's circle back
>begin with impartial morality
>end with small group of people deciding what quadrillions of hypothetical beings would want
>beautiful arc
>meanwhile actual humans keep doing annoying things
>voting wrong
>having parochial attachments
>loving family more than strangers
>caring about local community
>getting upset when told their suffering is cosmically negligible
>evolutionary biases everywhere
>explain that moral intuition cannot be trusted
>except intuition that future digital people count
>and intuition that extinction is uniquely bad
>and intuition that our probability estimates are sane
>and intuition that our institutional choices improve the future
>those intuitions survived peer review
>someone donates $5k to local homeless shelter
>inefficient
>could have funded 0.0000000000003% of an AI governance researcher
>think of all the simulated people you just killed
>okay maybe don't phrase it that way publicly
>PR team says "future generations deserve a voice"
>much better
>journalist asks what longtermism means
>say "future people matter"
>everyone agrees
>great
>journalist asks what follows from that
>well technically we should redirect enormous resources toward low-probability interventions affecting astronomical futures
>journalist raises eyebrow
>return to "future people matter"
>motte has entered the chat
>critic: of course future people matter
>me: glad we agree
>critic: I don't agree that your institute knows how to help them
>me: why do you hate our grandchildren
>eventually notice uncomfortable implication
>if future value dominates everything
>then helping people today mostly matters through effects on future
>education matters because future institutions
>health matters because future productivity
>democracy matters because future trajectory
>human beings slowly become instrumental variables in their own moral philosophy
>see starving child
>feel compassion
>check spreadsheet
>child's direct welfare contribution negligible
>but perhaps childhood nutrition improves national institutional quality
>compassion restored
>tell myself this is impartial altruism
>one day assistant asks obvious question
>"how do you know your intervention actually improves the far future?"
>silence
>open spreadsheet
>increase column width
>add confidence interval
>assistant asks again
>"no, I mean how do you know the sign is positive?"
>stare into cosmic light cone
>10^50 people staring back
>none of them exist
>none of them can tell me
>none of them can falsify my assumptions
>realize I have invented the perfect constituency
>infinitely important
>completely silent
>and always represented by me
Developing LLM based software hits different when you're working on a feature and you realize "Every time a user hits this button, it costs 20 cents."
Mechanical Dirk retweeted
My book, Reinforcement Learning from Human Feedback is done!
This is the book I wish I had when learning to fine-tune, align, & now post-train models since ChatGPT. The resource has been built by me finding time to study and document the fundamentals on nights and weekends since 2024.
Transferring as much of the intuitions of building Olmo as I possibly can in the book format.
The book is launching with an over 10 hour, full course with slidedecks, functional code for the training chapters, an example model completions library, and of course the free online web version.
Physical orders from Manning will ship in 1-2 weeks, and Amazon a week or so after. Thanks for your support!
Mechanical Dirk retweeted
I'm hiring Senior Research Engineers. Come build open-source vision-language models from zero to hero: pretraining, mid-training, post-training, RL, the whole pipeline.
job-boards.greenhouse.io/the…
Mechanical Dirk retweeted
Seven new models launching at Build: let’s go!
Reasoning. Code. Image. Transcribe. Voice.
Built from scratch on a clean data lineage, designed for efficiency, working seamlessly as a family of models
Thread 🧵
#MSBuild
ALT Graphic titled ‘Microsoft AI: 7 New Models’ showing icons for Image, Transcribe, Thinking, Voice, and Code models in a grid
This was a very cool project, and much of the work informed what we did for Olmo 3 long context. I wish we could have had the full results _before_ we started training...
New paper! allenai.org/papers/olmpool
This tackles a puzzle we found during the training of Olmo 3: how could two models with nearly identical short-context performance (and trained on the same data!) behave completely differently after long context extension?
Mechanical Dirk retweeted
(as seems to be the case with every release, Tyler saved our butts on post-training with some kernel wizardry)
We're releasing OLMo Hybrid. The core idea is that hybrid models (transformer attention + linear RNN layers) are strictly more expressive than either architecture alone, and that expressivity gap translates to better pretraining scaling. In a comparison to OLMo 3, we see 49% fewer tokens to match the same MMLU accuracy. Fully open model artifacts and technical report.
Mechanical Dirk retweeted
This plot undersells how much of a compute multiplier Olmo Hybrid is: 2x compute multiplier on many downstream tasks (and solid LC performance!!!)
I have the whole house full of Google Home devices. Couldn't wait for Gemini integration. But over the last few weeks we have reached a quality threshold where I don't even ask it to do things anymore, because I have no longer any expectation that it will work. @madeby
Not to mention the graveyard of features that worked when I bought this shit and no longer does. Listening to podcasts? Thermostat controls? Chaining commands with "and"? Executing tasks after a delay? All gone.
@Google is sitting on a gold mine with Google Home and Gemini, and they are bungling it so incredibly hard right now it's breathtaking.