Rustem Feyzkhanov

[Alex Ratner @ajratner]

The @SnorkelAI team did an audit of these findings using our internal QC platform.

Tl/dr: We agree with many of the findings, but ~55% of the fixes from @preferencemodel had their own issues with reward hacking or over-prescriptiveness.

Additionally: ~40% of the issues found were already tracked in the upstream repo - including 21 tasks with outstanding fix PRs!

Building datasets like this - and maintaining ambitious open benchmarks around them - is hard. We believe it requires open, collaborative review with community maintainers and relevant domain experts to get right.

@terminalbench was one of the first community efforts to build an open, continuous benchmark, and drive a vision of how OSS contributes to the rapidly advancing frontier of AI. We hope that the community will invest in working collaboratively with benchmark maintainers and task authors to keep this vision alive.

---

Some takeaways + details of our preliminary audit:

(1) We confirm 100% of the reward hacking / verification gap findings using our QC system. However, we find that 17/44 (39%) of the corrected tasks are just as or more reward hackable. Additionally, we don’t yet see evidence that these verification gaps meaningfully affected benchmark rollouts within CIs, though our analysis here is still ongoing.

(2) Our preliminary review raises questions about the underspecification findings for 22/41 (54%) of the flagged tasks, including 7 explicitly marked by task authors as intentional design choices. These cases warrant expert peer review and adjudication: balancing over/under-specification is a critical and tricky design decision - much of human intelligence concerns dealing correctly with reasonable and realistic underspecification - and we believe these corrections made to the tasks are likely too over-specified, reducing realism and difficulty relative to spec.

(3) Around 40% of the issues found were already tracked in the upstream repo (including 21 tasks with outstanding fix PRs)! OSS projects require community efforts to correct identified open issues, and ideally audits like these are done collaboratively with task authors and maintainers directly, in the traditional community spirit that has kept OSS alive for decades.

[vincent sunn chen @vincentsunnchen]

Today, we’re expanding @SnorkelAI’s Open Benchmarks Grants by 10x to a $30M commitment:
- Funding a more diverse, robust, and continuously-updated ecosystem of open benchmarks
- Launching the Open Benchmarks Red Team to continually test and strengthen them
- Introducing the Snorkel Research Fellowship to support independent contributors developing new evaluation methods

Since our launch earlier this year, we’ve been humbled to partner with the teams behind Terminal-Bench, TB-Science, ARC-AGI-3, OSWorld 2.0, Agents’ Last Exam, Continual Learning Bench, Senior SWE-Bench, SlopCodeBench, and more. Alongside these teams, we’ve helped design evaluation methodologies, develop task construction pipelines, and scale quality control. OBG-funded benchmarks have appeared on the latest model cards from every major frontier lab; have helped to index and measure frontier progress and alignment; and have contributed to guiding the frontier of AI development.

Benchmarks have always been guideposts for—and drivers of— AI progress: setting a metric and then “hillclimbing” against it is the core of how AI works! But when benchmarks fall behind frontier capabilities, our ability to evaluate and align these systems does as well. They become too simple, static, or correlated, and the risk of “benchmaxxing” increases.

We need more diverse, robust, and continuously-updated benchmarks, developed by a broader ecosystem of researchers and domain leaders.
We’re incredibly excited to accelerate frontier evaluation with Open Benchmarks Grants— raising the ambition and investment in robust, open, and independent measurement. Read our full announcement and apply here: https://t.co/E5kfZQ0Im5

[Rustem Feyzkhanov @Ryfeus]

@ajratner @SnorkelAI @insightpartners @S32_VC Exciting news!

[Rustem Feyzkhanov @Ryfeus]

Excited to be part of the team that made it possible to scale Data 2.0 platform. We are hiring so now is the best time to join!

[Alex Ratner @ajratner]

I'm excited to announce @SnorkelAI's $350M Series E at $3.5B, led by @insightpartners and @S32_VC.

We've grown 18x+ in the last 12 months since launching our Data-as-a-Service offering, passing $375M ARRR this week.

As AI advances to superhuman capabilities, AI data & environment development must advance with it - and basic staffing and crowdsourcing approaches are not enough.

AI progress now requires deep research and technology work that combines human expertise with specialized AI in compounding ways. @SnorkelAI is building the RSI data engine and frontier data lab for this next phase.

We're honored to have the support of existing investors Addition, @lightspeedvp, @GreylockVC, @GVteam, P7, Factory, @WellsFargo, Walden Catalyst Ventures, and new investors @ThirdPointLLC, @MarchCPs, @BlumbergCapital, @AllegisCapital, @Frontlinevc, and @standard_vc.

–

@SnorkelAI started as a research project a decade ago at @StanfordAILab.

Our thesis was simple: AI progress would become increasingly data-centric – and therefore data development should be studied as a true research and technology problem, not just a staffing and crowdsourcing one.

Today, as AI capabilities verge on superhuman, building the data and environments to safely measure and train AI is becoming too hard for even the smartest human experts to do alone.

Only humans and AI agents, collaborating together in compounding ways, can meet the accelerating needs of the frontier, and keep humans in the driver’s seat of AI progress for decades to come.

At @SnorkelAI, we are building the data lab to define the shape of this new “Data 2.0” frontier, and the new paradigms of human-computer interaction needed to advance it.

Our key focus is building the RSI engine for data, where specialized AI models accelerate and improve human expert output, and in turn, scaled human supervision is used to continuously evaluate and improve these models – creating a powerful compounding loop to keep pace with an accelerating RSI frontier.

With this round of funding, we are also doubling down on our commitments to support data development for open benchmarking and evaluation (more news here soon!); an increasingly diverse ecosystem of general and specialized intelligence; and a path to safe, well-aligned AI built on robust training and evaluation data.

Data development will guide and drive the next stages of AI – and must do so in a human-centric, AI accelerated, open, diverse, and safe way. We are excited to support this mission in the next decade of research ahead at @SnorkelAI.

More thoughts here: https://t.co/Dzk6olqmAc

[Alex Ratner @ajratner]

Agentic AI will transform every enterprise–but only if agents are trusted experts.

The key: Evaluation & tuning on specialized, expert data.

I’m excited to announce two new products to support this–@SnorkelAI Evaluate & Expert Data-as-a-Service–along w/ our $100M Series D!

---

Snorkel Evaluate is our new data-centric agentic AI evaluation platform for specialized, mission-critical enterprise settings where vibe checks and out-of-the-box metrics driven by simple LLM prompts are not enough.

Snorkel Expert Data-as-a-Service is our white glove service for expert-level AI datasets, powering frontier LLM developers in areas like expert knowledge, reasoning, agentic action and tool use, and more!

Both built on top of @SnorkelAI’s Data Development Platform, using our programmatic technology to drive higher-quality expert data, faster– for getting specialized AI to real production value.

If you’re building enterprise AI and want to partner around the key ingredient in AI today–the data–book a demo and let's talk! https://t.co/w0J8izpn8p

Finally, see thread for details on 🧵👇
- 📽️ A walkthrough of Snorkel Evaluate and Expert Data-as-a-Service on an agentic AI enterprise task
- 📅 An upcoming event on Enterprise Agentic AI with innovators from @Accenture @BNY @Comcast @Stanford @QBE & others
- 📊 An upcoming series of benchmark datasets and model artifact releases

👀 Want early access to the full agentic AI dataset? Retweet this post and we'll send you the link!

[Embedded Vision Summit @EmbVisionSummit]

Looking to efficiently scale machine learning solutions from research to production? Join us to hear Rustem Feyzkhanov, Machine Learning Engineer at @instrumentalinc, present his talk at the Embedded Vision Summit, May 20-22:

https://t.co/AiPJNvOfrm

[Rustem Feyzkhanov @Ryfeus]

Hey folks. Happy to share my blog post on how to use AWS Lambda SnapStart for LLM inference https://t.co/EFlygMTXzC
Quick summary: exciting, but limited to frameworks which are both small (250MB zip limit) and can load the models from memory, not hard disk (512MB limit).

[Amazon Web Services @awscloud]

Dive deep into #generativeAI at #AWSreInvent. ☁️ ⚙️ 🤖

AWS #MachineLearning Hero @ryfeus shares an Attendee Guide that highlights the re:Invent sessions that will help you navigate the rapidly evolving landscape of generative models.

Learn more. 👉 https://t.co/rhMz89o45r

[Chris Williams @mistwire]

Day 1 of the @awscloud Hero Summit (super NDA so no screenshots 😅)
A LOT of brainpower in this room 🤯😳

[Ken Collins @metaskills]

Bus 🚌 trip!!!

[AWS Developers @awsdevelopers]

Let's scale, community! #AWSHero @Ryfeus has a new article on how to create a Scalable Cloud inference endpoint. https://t.co/FDnN73z3rX

[Instrumental @instrumentalinc]

Excited to announce that @AlonBenShoshan has joined Instrumental as our Chief Business Officer as we look to enable the world's most admired electronics brands to build better products.

Check out the news here! https://t.co/6dBZ4lNVzN

[Ken Collins @metaskills]

🎁 HAPPY HOLIDAYS!!! 🎁

New Lamby v4 site using @Docusaurus is now live. What do y’all think?

[Hiromi Ito🌞🌏���︎Que será, será @hiro_baila]

Yesterday was another great night🤩🤟It was great fun to see so many friends again and meet new ones. We all had a lot of fun talking about our countries and keynotes🥳#reInvent2022
@colors4ideas @franciscoed @vfcarida @Ryfeus @antonbabenko

[Farrah C @FarrahC32]

If you are attending @AWSreInvent I hope you find time to stop by the Modern Applications and Open Source Zone. Starting at 9 am tomorrow in the Ventian (Third Floor, East Alcove).
Find us in the session catalog (ACT083) or at: https://t.co/bPxvAxMZSV
(cont'd)

[Ross Barich @rossbarich]

Attending AWS #reInvent? Don't miss the #AWSCommunity session track, led by AWS Heroes!

⭐️ Favorite in the re:Invent session guide now, & grab a seat when reserved seating launches October 11, 10AM PDT:

https://t.co/ovcDIf0ek4

[Mike Fiedler, Code Gardener @mikefiedler]

For those looking to squeeze a few more milliseconds out of Python Lambdas, here's a tip that can help! https://t.co/FGQiry3TAc

[Anton Babenko 🇺🇦 @antonbabenko]

Let's gather as many people as possible and have interesting face-to-face discussions in #Paris on the 11th of October!

Join us if you are curious about #DevOps, #Infrastructure as Code, #Terraform, #Pulumi, #CDKs, and everything in between!

https://t.co/NRQKjP52ry

[Adam Selipsky @aselipsky]

Best part of #reInvent is connecting with customers, partners, and teammates!

[Vlad Ionescu (he/him) @iamvlaaaaaaad]

How should you run new (production) containers on AWS?

I wanted to write a longer blog post about this, but 🤷. So, here's a flowchart!

Because I know people will "But actually", there's a second version of the flowchart with some notes.