AI Agents Product Lead 11x AI
Joined December 2015
- Tweets376
- Following387
- Followers138
- Likes3.8K
Sachi Angle retweeted
Literally unlocking net new capabilities in multimodal memory with @GraphonAI @typesafeai
I mapped an entire movie by mood for about 5 cents (Jev @typesafeai + @GraphonAI).
Moods like: Funny, Scary, Thrilling, Suspenseful etc.
Around 7,000 AI analyses for roughly $0.05 and 7 seconds.
The demo is pretty simple:
First, Graphon maps the movie.
We ask Graphon about 40 questions designed to cover what's happening throughout the video. Graphon returns descriptions of the scenes, along with citations to the exact moments in the movie.
We use those questions to build a picture of what's happening across the entire movie.
Then we send those scene descriptions to Jev from @TypeSafeai
Jev looks at each scene and judges the probability that it is funny, scary, suspenseful, exciting, calm, or whatever other categories we want to analyze.
That's where we do around 7,000 analyses in parallel.
The result is basically a mood map of the entire movie.
Now imagine you're a movie studio, video editor, or creator.
You could say:
“Show me every scary moment.”
“Find all the car chases.”
“Give me the funniest 30 seconds.”
“Create a compilation of the most suspenseful moments.”
And jump directly to those exact scenes.
Today, doing this at scale is surprisingly complicated. Someone has to watch, tag, classify, scrub through footage, and manually find all of these moments.
With Graphon + Jev, you can start doing it automatically.
And the crazy part?
Those ~7,000 mood analyses cost about 5 cents.
This is what I think multimodal search starts to look like when you combine understanding, citations, and reasoning across an entire video.
@CompleteSkeptic what do you think?
Sachi Angle retweeted
"@getbluejay_ai is not a tool we use. They are partners." That's Muizz Matemilola at @11x_official. Our case study with them is out today.
Some background on 11x: they've raised over $70M from a16z and Benchmark, and their agent Julian handles inbound demand for go-to-market teams at companies like Checkr and Canibuild. Julian talks to real prospects, in a customer's name, in the wild.
Before Bluejay, validating a single change meant calling the agent over and over, one conversation at a time, and edge case coverage stopped at what the team could imagine. Every gap was a risk in front of a real prospect.
Here's why 11x chose to #GoBlue:
1/ Thousands of realistic simulations before an agent ever takes a live call
2/ Automated red teaming graded against OWASP, MLCommons AILuminate, and MITRE ATLAS
3/ Living eval suites that evolve with each customer instead of a checklist that only grows
4/ A self-healing loop where agents improve agents until quality clears the bar
5/ A true partner behind it all. They came to us with one problem, we solved it together, they found more, we solved those too
So far that's saved them over 1,100 hours of manual QA. More importantly, their customers deploy faster because they can see the agent already survived scenarios they hadn't thought of themselves.
Thank you to @sachiangle, Muizz Matemilola, @flanciscoz, @aaronanan, Santiago Gutierrez, and the entire 11x team for sharing their story.
Full case study linked in the comments.
Sachi Angle retweeted
Excited to see @11x_official featured in @a16z's AI Application Spending Report!
📷 Big thanks to the @a16z team for the thoughtful coverage on where enterprise AI is heading.
If you’re curious about how to supercharge your GTM teams with AI, reach out!
Sachi Angle retweeted
oh, you’re into AI? then tell me about…
GPT, Claude, Gemini, LLaMA, Mistral, Mixtral, Grok, DeepSeek
Kimi, Qwen, Ernie, Yi, Gemma, Zephyr, Falcon, Phi
Hermes, Perplexity, Sora, Runway, Pika, Luma, CharacterAI
Alpaca, Vicuna, Orca, Dolphin, Mosaic, PanGu, Jurassic, Chinchilla
BLOOM, Pythia, Megatron, Nemotron, PaLM, LaMDA, Gopher, Flamingo
Kosmos, BLIP, LLaVA, MiniGPT, Whisper, Muse, Titan, Granite
BabyAGI, AutoGPT, SuperAGI, AgentGPT, LangChain, LlamaIndex, LangGraph, CrewAI
DSPy, Agno, AutoGen, AgentFlow, SmolAgents, SemanticKernel, PromptLayer, PromptPerfect
RAG, GraphRAG, Guardrails, TrustLLM, Together, Anyscale, Fireworks, Replicate
Baseten, Modal, VLLM, TensorRT, ONNX, Triton, Ollama, LMStudio
GPTQ, AWQ, GGUF, ExLlama, Tabby, LoRA, QLoRA, HuggingFace
PEFT, RLHF, Replit, Copilot, Jules, Tabnine, Cursor, Aider
Devin, Vibe, Base44, Bolt, Lovable, Windsurf, Suno, GenSpark
v0, Manus, Sesame, Soundful, ElevenLabs, VoiceMaker, Pano, Pi
NVIDIA, Groq, Cerebras, SambaNova, Graphcore, Tenstorrent, Habana, Mythic
LightOn, Optimus, Figure, Digit, Spot, Atlas, BostonDynamics, Unitree
Sachi Angle retweeted
I think context engineering is going to stick - unlike "prompt engineering" it has an inferred definition that's much closer to the intended meaning, which is to carefully and skillfully construct the right context to get great results from LLMs
+1 for "context engineering" over "prompt engineering".
People associate prompts with short task descriptions you'd give an LLM in your day-to-day use. When in every industrial-strength LLM app, context engineering is the delicate art and science of filling the context window with just the right information for the next step. Science because doing this right involves task descriptions and explanations, few shot examples, RAG, related (possibly multimodal) data, tools, state and history, compacting... Too little or of the wrong form and the LLM doesn't have the right context for optimal performance. Too much or too irrelevant and the LLM costs might go up and performance might come down. Doing this well is highly non-trivial. And art because of the guiding intuition around LLM psychology of people spirits.
On top of context engineering itself, an LLM app has to:
- break up problems just right into control flows
- pack the context windows just right
- dispatch calls to LLMs of the right kind and capability
- handle generation-verification UIUX flows
- a lot more - guardrails, security, evals, parallelism, prefetching, ...
So context engineering is just one small piece of an emerging thick layer of non-trivial software that coordinates individual LLM calls (and a lot more) into full LLM apps. The term "ChatGPT wrapper" is tired and really, really wrong.
Sachi Angle retweeted
I think there’s a future where people are like “let’s do a AI-less meeting” similar to how we do camera-off meetings
