@bjamin_ar

easily distracted by bikes // mts @adaption_ai // @neo @middlebury

 /sf
benjamin.ar
Joined July 2018
  • Tweets2.2K
  • Following440
  • Followers1.2K
201Media
benjamin ar retweeted
adaption
@adaption_aiOct 15
Introducing Visualize, the first release in Adaption’s Adaptive Interfaces. Turn your insights into actions that shape your data and intelligence with Adaption’s full AI system. See it. Shape it. Own it.
18
30
10
158
16,826
Checkout the recent work from @weiyinko_ml and @sarahookr! We've got a lot of headroom for training improvements with data modifications. Making this scale across diverse domains is an impressive feat.
adaption
@adaption_aiOct 8
New research from Adaption presents adaptive agentic checklist generation. An agent builds a quality checklist from past AutoScientist runs. The checklist scores and removes bad training data. Model performance improves every run.
6
319
benjamin ar retweeted
adaption
@adaption_aiOct 8
New research from Adaption presents adaptive agentic checklist generation. An agent builds a quality checklist from past AutoScientist runs. The checklist scores and removes bad training data. Model performance improves every run.
5
12
4
48
4,443
benjamin ar retweeted
Shivalika Singh@singhshiviiiOct 1
We recently launched Invent-a-dataset that allows creating post-training datasets using just a natural language description 📖 🖊️ Today we share an extensive technical report evaluating Invent API against using frontier proprietary and open-weight LLMs for data generation ✨🧵
4
26
3
146
11,887
benjamin ar retweeted
adaption
@adaption_aiOct 1
The biggest hurdle in frontier AI is curating high-quality, diverse data. What happens if you need to post-train a model for a new capability, but have zero data? Today, we release the technical report for Invent a Dataset.
31
44
18
246
68,047
high quality work towards a zero data regime from the rest of the team!
adaption
@adaption_aiOct 1
Replying to @adaption_ai
Benchmark tests across 8 task types show Invent significantly outperforms top frontier APIs (Claude Opus 5, GPT-5.6, Gemini 3.1 Pro, DeepSeek V4 Pro, and GLM-5.3) with +17% higher quality and +19% greater sample diversity. Invent API's diversity advantage widens as dataset size grows, reaching a +37% diversity lead at 20K samples with 0.0% duplicates.
1
9
370
benjamin ar retweeted
adaption
@adaption_aiSep 15
Last week we introduced Invent a Dataset. Describe what you need. Hit go. Today, we’re making it even easier with the Invent API. A few lines of code. AI ready training datasets in minutes.
10
32
6
151
61,602
I'm a member of emoji staff at adaption
adaption
@adaption_aiSep 14
Meet Grady. Tardigrades are nearly indestructible. They adapt to any environment, from the ocean floor to the vacuum of outer space. Learn the story of how Grady became our Adaption mascot.
2
8
724
benjamin ar retweeted
adaption
@adaption_aiSep 14
Meet Grady. Tardigrades are nearly indestructible. They adapt to any environment, from the ocean floor to the vacuum of outer space. Learn the story of how Grady became our Adaption mascot.
12
12
3
53
24,411
keep @dhruvrnaik away from your harness if you want to keep tokenmaxxing
adaption
@adaption_aiSep 9
New research from @sudip_r0y and @dhruvrnaik. Harness optimization moved criterion pass rate from 67.10% to 85.92% on @harvey Legal Agent Benchmark. Post-training pushed it to 88.03%. Owning your intelligence means improving the model and the harness around it.
1
1
158
benjamin ar retweeted
adaption
@adaption_aiSep 9
New research from @sudip_r0y and @dhruvrnaik. Harness optimization moved criterion pass rate from 67.10% to 85.92% on @harvey Legal Agent Benchmark. Post-training pushed it to 88.03%. Owning your intelligence means improving the model and the harness around it.
5
16
5
53
7,094
benjamin ar retweeted
adaption
@adaption_aiSep 3
Introducing Invent a Dataset. Describe the dataset you need. Get a structured, training-ready dataset back. No existing data needed. Dataset creation used to start with collection. Now it starts with specification.
33
88
23
831
93,021
2FA has gotten out of hand
1
8
175
I don't know about y'all but i'd never played a version of this where you get interrogated after sharing your three statements @sarahookr was quick to change that 😅
adaption
@adaption_aiAug 19
Co-founders @sarahookr and @sudip_r0y played Two Truths and a Lie with the team. Turns out the people behind the research have a few stories of their own. Can you spot the lie?
2
5
1,234
some great new work from @bargav025 and @sudip_r0y improving on both cost and latency
adaption
@adaption_aiAug 13
The cost of remembering should not scale with how much there is to remember. New research from @bargav025 and @sudip_r0y: 45% faster on long conversations, 97% fewer tokens processed.
6
201
benjamin ar retweeted
adaption
@adaption_aiAug 12
First episode of Field Notes, our new monthly series on the leaders shaping AI. @sudip_r0y, Adaption Co-founder, and @mgalle, post-training lead at @poolsideai discuss the last 5% of reliability and what it actually costs.
8
12
3
34
10,041
it's fun to build AI for the rest of the world
adaption
@adaption_aiAug 4
Better training data. Faster training cycles. Any language, any domain. @AISingapore used Adaption to enhance dataset quality and expand training dataset size via localization across five low-resource Southeast Asian languages.
8
438
benjamin ar retweeted
adaption
@adaption_aiAug 4
Better training data. Faster training cycles. Any language, any domain. @AISingapore used Adaption to enhance dataset quality and expand training dataset size via localization across five low-resource Southeast Asian languages.
6
27
5
60
11,986
will and this team are incredible at what they do. it’s a mandate that they will do great things.
Will Hoppin
@willyhoppsAug 3
hello everyone, announcing something BIG soon...
1
4
256
benjamin ar retweeted
Will Hoppin
@willyhoppsAug 3
hello everyone, announcing something BIG soon...
109
39
28
359
56,625
Load more