vp of ai research @beaconholdings prev ceo @haizelabs
sf
Joined May 2013
- Tweets2.8K
- Following1.3K
- Followers4.7K
- Likes14.9K
Pinned Tweet
I am thrilled to announce that @beaconholdings has acquired @haizelabs, with me joining as VP of AI Research.
We started Haize in 2024 to enable anyone to build reliable and safe AI. Through our red-teaming and safeguards work with the frontier labs; our observability, guardrail, and evaluation platform serving the world’s largest enterprises; and our pro bono SMB work, any customer could build AI they trusted with Haize.
Joining Beacon lets us deliver the same trustworthy AI to those who need it most, and those most overlooked by Silicon Valley: the essential Main Street businesses the real world depends upon.
Transitioning these essential businesses through the AI revolution is one of the most consequential responsible AI problems of our time. We couldn’t be more honored to tackle it with Beacon.
Thank you to our customers, investors, and team for the journey of a lifetime. And thank you to Nilam, Goutham, Mark, and the Beacon team for the trust and opportunity.
It’s time to get to work.
dachi is the closest thing i've felt to a natural sounding human to date!
so excited for the world to experience dachi now 🫂
Frontier models are RL'ed to excel at tasks but have become annoying to talk to.
So we built Tomo’s first model, Tomo Dachi 1.0
Users on Dachi retain longer, complain less, and engage more, beating Opus 5.5 and 6.1 Sol in prod while maintaining strong agentic ability. [1/n]
3
15
1,786
this is very much not a world i wish humanity to inhabit
i have been thinking
really,
really,
really
hard
about how to avoid this—currently likely—future governed by hyper-rentiers that asymmetrically benefit the elite
The benefits of abundant intelligence will accrue far outside of San Fransisco.
Spent the last year writing about capital, sovereignty, and the coming reordering of economic power.
THE COMMANDING HEIGHTS
Commanding-Heights.com
the Great Transformation is upon us
congrats team!
Today, we’re introducing the next chapter of Concurrence and the world’s greatest leap toward Superintelligent Care.
When I was 13 I lost my mom to cancer that should’ve been caught earlier.
We’re building a system that sees every patient’s future in time to change the outcome
3
6
1,188
now what if we tried to do this for real world / businesses..?
lmk @rosstaylor90 :)
Introducing Reinforcement Learning from Collective Rewards (RLCR).
RLCR is a recipe that trains language models as agent organisations, which share a message board and workspace. We use it to train Avalon-M1-Mini (120B) and study its behaviour.
We find a number of interesting results, including:
🛠️ Learnt organisational hierarchies and divisions of labour
🕵️ Increased collective verification, through hires of verifiers and testers
👥 The rise of collective thinking, where agents think more about other agents and how to help them
We believe RLCR can provide a laboratory for studying collective intelligence and multi-agent safety.
Blog post and technical report 👇
1
11
2,158
If you spend too much time in the Valley, you lose touch with reality and the broader human condition
It's necessary to have your boots on the ground to uncover what is true
I had the please of speaking with @the_logic about what we are building at @beaconholdings
Thank you to @muradhem for taking the time to get to get to know us
1
14
1,473
this is the future…
Spent years watching models ace a single image and completely whiff on anything that moves or changes angle. Today we’re fixing that.
Try Glimpse now, our first visual thinking model with competitive frontier performance compared to Astra on key public visual reasoning benchmarks.
10
1,344
high quality benchmark from a high quality team!
Introducing DeepRSI, a new industry standard for measuring recursive self-improvement.
On preexisting benchmarks, tasks are contributor-based and short-horizon (~12 hrs). In DeepRSI, we push agents to work 48 hours on ML tasks actually derived from ongoing proprietary research.
4
1,196
personalized, private intelligence is so much more than it appears on the surface
no doubt it will be useful
but it is also a viable path towards solving safety & alignment in the long run
congrats Zain and team!
Introducing Core, The Personal Al Computer
The most powerful technology humanity has built should belong to you
Order today: ghost.ai
1
20
1,996
my mcmc goat
so cool to see classic ideas to be re-brought to life
SAMPLING + SFT ARE ALL YOU NEED
SFT is not dead! 🥳
We found a way to make SFT rival current prevailing posttraining methods, often generalizing better and forgetting less than RL and OPSD. 🤯
Following our prior work on reasoning with sampling, we now introduce sampling to the posttraining stack.
1/n
2
31
8,061
formidable people radically expand your conception of what is possible—for yourself and for humanity
i don’t think formidability is innate!
some thoughts on how formidability is learned:
leonardtang.me/blog/on-formi…
10
952
As @nilamg points out—for the people running schools, factories, and oil rigs, trust and safety is a non-negotiable for adoption
Transforming models into systems they can trust is central to our work @beaconholdings
Our Founder & CEO @nilamg sat down with @Bloomberg to talk about the real barrier to AI adoption: trust.
Nilam shares more about the model we're building at Beacon, think Berkshire Hathaway rebuilt for an AI-native world, as the first centralized AI holding company.
bloomberg.com/news/videos/20…
13
1,501
A generation of small business owners will retire in the Silver Tsunami; and a lifetime of know-how will otherwise retire with them
Company World Models (CWMs) from @beaconholdings absorb these owner-operator's expertise while they're still at the helm
to immortalize their life's work
the everyman,
essential to keeping the lights of the world on,
deserves a say in how intelligence is shaped,
deserves intelligence that works for them
@nilamg and I went on @tbpn to share more on who we're serving and what we're building @beaconholdings
Essential technology behind essential work.
Focus FS, a @beaconholdings portfolio company, helps mine rescue teams share critical information.
We’re building with Focus FS to help miners learn from historical incidents to robustify their safety program.
Watch the full breakdown of Focus FS technology from Dräger Global @dragerglobal here:
youtu.be/Z6Z2cUsj5R8
3
504
#1 today 🥺🥺
thanks y'all 🫶
I am thrilled to announce that @beaconholdings has acquired @haizelabs, with me joining as VP of AI Research.
We started Haize in 2024 to enable anyone to build reliable and safe AI. Through our red-teaming and safeguards work with the frontier labs; our observability, guardrail, and evaluation platform serving the world’s largest enterprises; and our pro bono SMB work, any customer could build AI they trusted with Haize.
Joining Beacon lets us deliver the same trustworthy AI to those who need it most, and those most overlooked by Silicon Valley: the essential Main Street businesses the real world depends upon.
Transitioning these essential businesses through the AI revolution is one of the most consequential responsible AI problems of our time. We couldn’t be more honored to tackle it with Beacon.
Thank you to our customers, investors, and team for the journey of a lifetime. And thank you to Nilam, Goutham, Mark, and the Beacon team for the trust and opportunity.
It’s time to get to work.