CEO/Founder of @joinhandshake
San Francisco, CA
- Tweets 3,005
- Following 2,647
- Followers 7,359
- Likes 10,002
Garrett Lord retweeted
Our frontier work in model security and alignment is what I’m most excited to bring to enterprise and a major bottleneck for sovereign SI
Meta’s Muse Spark 1.3 Evaluation Report describes the human red teaming Handshake’s team did before Muse launched (section 3.1.4.2). Pre-release access, jailbreaks found through the product, full trajectories and analysis sent back, mitigations improved before release.
Report: research.meta.ai/static/muse…
1
8
606
Garrett Lord retweeted
Claude scored 100% on structured accounting tasks, but succeeded in only 12% of full banking assignments. Being good at individual financial tasks is different from delivering actual financial work inside an enterprise.
@airstreet's 2026 State of AI Report highlights this gap using our ATLAS-Finance benchmark.
In 2026, AI delivered outsized value to software engineers and vibe coders. 2027 will be a breakout year for financial agents, driven by high-fidelity evaluations like ATLAS-Finance and RL environments that train agents to accurately produce high-value deliverables.
2
3
17
1,733
Garrett Lord retweeted
Meta’s Muse Spark 1.3 Evaluation Report describes the human red teaming Handshake’s team did before Muse launched (section 3.1.4.2). Pre-release access, jailbreaks found through the product, full trajectories and analysis sent back, mitigations improved before release.
Report: research.meta.ai/static/muse…
2
5
12
1,584
Garrett Lord retweeted
How Garrett Lord decides which problems get his time at Handshake:
"I like to think about problems of red, yellow, green. What's red today but will be green in six months? I try to not focus on those problems.
A lot of the team wants to flag issues to me where there's a fire, and I'm like, red, yellow, green, is it going to be fine in a quarter? Is it going to be fine in two quarters? If it's going to be fine, if you guys have the ball, I'm not going to spend my time there.
I'm going to try to focus on things that are red today that will be red in the future."
- @GarrettLord
1
12
303
7,768
Garrett Lord retweeted
"Owning your own intelligence" = caring enough to actually spend real time, attention, and resources on building great evals for every part of your business
1
1
6
677
Garrett Lord retweeted
Evals for the agents
Evals for the agents monitoring the agents
Evals for the evals for the agents and the agents monitoring the agents
4
2
22
1,025
Garrett Lord retweeted
Exciting to see @joinHandshake alongside some of the teams advancing the next generation of AI.
Proud to be part of @nvidia's Nemotron ecosystem!
4
4
19
1,956
Garrett Lord retweeted
Congratulations to @reflection_ai on Beam, released today with a full safety report.
Handshake is named in the report as a source for the conversations Beam was tested against, and among the partners who tested the model before release.
Safety testing is expert work. Before a model ships it has to face the requests that matter: self-harm, child safety, cyber offense, fraud. The hard cases are not the obvious ones. A nurse or a researcher can ask something that sounds dangerous and isn’t. Someone has to write that case and decide where the safe answer ends.
Our people wrote those cases for Beam and judged every turn. Making AI safe takes the same expert judgment as making it useful, and we do both.
Report: reflection.ai/beam-safety-re…
Beam Safety Technical Report — Reflection
Make intelligence open and accessible to all
reflection.ai
Beam is now available across platforms, with open weights under the Apache 2.0 license - in collaboration with an ecosystem of 40+ partners.
We're releasing:
- Weights, tokenizer, and reference implementation
- BF16, FP8 and NVFP4 weights and a DFlash2 drafter
- The open source harness we built and trained Beam on: Mirror CLI
Weights and model card on Hugging Face: huggingface.co/reflection
1
15
2,134
congrats @AndrewDai!
1
17
1,288
Garrett Lord retweeted
In terms of performance-vs-cost, @OpenAI's GPT-6.1 Sol & @GoogleDeepMind's Gemini-4 do represent big leaps in the computer vision capabilities of models in specialized scientific domains.
1
3
8
342
Garrett Lord retweeted
We doubled the accuracy of Nemotron 3.5 Super VL on visual scientific assay interpretation, fine-tuning it into a Pareto-optimal model.
Excited to build more specialized multimodal models from open weights + data we’ve created!
joinhandshake.com/research/a…
Meet NVIDIA Nemotron 3.5 Super VL
Our new open model combines reasoning and tool use with text, image, document, and video understanding for long-running AI agents.
120B total parameters with 12B active 🧵
1
2
8
508
Garrett Lord retweeted
In collaboration with @NVIDIAAI , we post-trained Nemotron 3.5 Super VL to understand and reason on scientific artifacts. We collected high quality data from industry experts, and measured the performance on Handshake's ATLAS-VIALS (life sciences) benchmark. Training doubled Nemotron’s accuracy on VIALS with performance gains that also transferred to related benchmarks; putting it on pareto frontier of cost/accuracy.
We at @joinHandshake are excited to build accurate, cost-effective AI, using post-trainable models that enterprises can securely deploy on their work. Read more here: joinhandshake.com/research/a…
7
15
23
765
Really proud of our team. We want to be the #1 platform Fellows trust.
Every payment is now broken down and every adjustment is explained. We’ll keep iterating as the industry evolves.
Mission: be the best place to learn AI skills, find and navigate jobs, and get paid for what you know.
joinhandshake.com/blog/our-t…
Fellows: your dashboard is changing so you can see what you earned, how it was calculated, and when it lands. You’ll see base pay, incentives, and adjustments by project, plus a reason for every change and payment status.
Rolling out across projects over the next few weeks.
Full update here: joinhandshake.com/blog/our-t…
1
31
1,688
