Anne Marie Droste

2.2K posts

Anne Marie Droste banner
Anne Marie Droste

Anne Marie Droste

@AmDroste

Co-Founder at Tacit Labs Evals & hocus pocus

San Francisco Se unió Haziran 2012
1.4K Siguiendo1.7K Seguidores
Anne Marie Droste retuiteado
Neal Amin
Neal Amin@neal_amin·
@bauer_lesavage @_DimensionCap biological data is heavily corrupted. Crystallography protein structures are the exception. Measuring nature is the real bottleneck to ai bio
English
1
0
0
124
Bauer LeSavage
Bauer LeSavage@bauer_lesavage·
The training data market has exploded for LLMs and bio foundation models are next. But biological data is extremely complex and requires a data generation playbook that prioritizes quality over immediate scale. @_DimensionCap Research article live now! research.dimensioncap.com/p/on-training-…
Bauer LeSavage tweet media
English
9
15
107
24.9K
Anne Marie Droste retuiteado
David
David@DavidSHolz·
dream harder
English
99
301
2.5K
265.8K
Anne Marie Droste retuiteado
Nicole Fitzgerald
Nicole Fitzgerald@ninklefitz·
all the @tacitlabsco brand work was done by my brother, Connor, who is infinitely cool and my favourite person to make stuff with. everyone should work w him! contact info below -
🌎@garbnzgh

@ninklefitz @tacitlabsco Congrats! Who did the brand work? It’s very nice

English
7
2
144
15.8K
Anne Marie Droste retuiteado
Karan Singhal
Karan Singhal@thekaransinghal·
One of the ways scaling test-time compute can benefit people most: have reasoning models think really hard about rare undiagnosed diseases. Today we’re sharing published evidence that this can work, in some of the most difficult pediatric cases!
OpenAI@OpenAI

Together with researchers at Boston Children’s Hospital and Harvard, we published a study in NEJM AI showing how o3 Deep Research helped clinicians revisit previously unsolved rare pediatric disease cases, and find answers for families who had waited years.

English
89
82
715
320.1K
Anne Marie Droste
Anne Marie Droste@AmDroste·
When I first met @ninklefitz, she convinced me that science is the next LLM research frontier, long before that became obvious. Together we started @tacitlabsco and set out to compress the 10 year, $1B verification loop in biopharma. Fast, abundant, reliable verifiers are the single biggest lever we have in meaningfully training and deploying models. I believe in a future where drug discovery and development is abundant; millions of agentic, zero person biotechs will explore therapeutic hypotheses autonomously, in parallel. We're hiring!
tacit labs@tacitlabsco

Introducing Tacit Labs: an applied research lab at the intersection of AI and biology.

English
10
3
58
12.5K
Nicole Fitzgerald
Nicole Fitzgerald@ninklefitz·
I've started a new company called @tacitlabsco, an applied research lab at the intersection of AI and biology. 1/n
English
53
32
540
110.5K
Anne Marie Droste
Anne Marie Droste@AmDroste·
@nicetomeetyu2 Ha! OAI are (and Ant the other way around) restricted from each others' API. OpenAI are planning on hosting the benchmark on an independent platform soon so we'll see Claude scores.
English
1
0
1
185
Anne Marie Droste
Anne Marie Droste@AmDroste·
I think that as evals become more realistic, they will increasingly rely on ground truth signal (ie raw biochemistry experimental data), rather than 'opinion', so yes. At the same time, there's a *lot* of human knowledge that we haven't captured yet, and deterministic verifiers force black and white answers where the reality is grey.
English
1
0
1
13