Nick Jalbert

MoTS @LaudeInstitute. Formerly @perplexity_ai.

[Laude Institute @LaudeInstitute]

Introducing Slingshots // FOUR: 14 projects across eight top institutions. Self-improving agent systems and harnesses, maximally-open models and training data, applications for research and decision-making – plus kernels and more. Say hello to the batch 🧵

[Nick Jalbert @NickAJalbert]

Reviewing generated code by hand sucks. Timely benchmark from the @github team! Excited RevEngine landed in the top spot, but looking forward to seeing this one saturate so I'm out of the diff mines for good.

[Laude Institute @LaudeInstitute]

When we released https://t.co/TyCu4paN4k, we got a spike of contributions (good problems!) but found ourselves knee-deep in PRs.

Luckily, Laude’s eng team has a lot of freedom to experiment, plus a tendency to favor DIY over off-the-shelf. @NickAJalbert, Justin Fiedler, @andykonwinski & team built and open-sourced RevEngine to hoist humans out of the loop during code review. It's since landed at the top of @GitHub’s ReviewBench leaderboard.

More on process and learnings in a blog soon. Let’s see if we can hold #1 until then 👀

[Nick Jalbert @NickAJalbert]

This one has been brewing for a while! Multiplayer was a big surprise--super fun. Definitely an under-explored interaction paradigm.

[Andy Konwinski @andykonwinski]

Introducing Headlong, an open source microharness for persistent agents: self-guided agents that think continuously.

Most agent harnesses are reactive: you send a task, the agent completes it, and then it sits frozen until the next request. Cron jobs and heartbeats wake it up to run a checklist and put it back to sleep.

A Headlong agent is never asleep. It keeps generating thoughts about whatever it decides is interesting, in a self-guided loop inspired by human inner monologue. Your message doesn't start a session. It's one more observation that lands in the agent's thought stream, and the agent decides if and when to reply.

Headlong is built on the idea of persistent agency: continuous inner thought generation between external interactions. The agent sets its own interests and priorities, comes up with its own projects, and sometimes pings you unprompted with progress.

To keep our prototype as simple and small as possible, we implemented Headlong as a microharness: a complete agent harness in under 10K lines of Bash, organized as a handful of small executables. It includes a loop that generates the next thought, shellm (a recursive language model written in Bash), a trajectory stored as a DAG of jsonl files, and context as a projection of that trajectory.

We've been running one Headlong agent internally at Laude for several weeks. The whole team talks to it over Slack and Telegram, and every conversation lands in its single stream of thought. It works in its own fork of Headlong and we've pulled over 50 of its commits into main.

One night, with nobody talking to it, it went back to check whether a recall process it had built was actually wired into its mind, found that it wasn't, diagnosed and fixed the bug, and verified the fix end to end. 48 minutes, no human asked for the fix or was in the loop at any point. Every step is a timestamped line in its log.

Things broke too, and we wrote those up. Background thinking costs us $1 to $2 an hour, our agent stopped its own service three times by accident, and self-delegation died on day one. Details in the post.

One line installs everything and starts an agent. Use a dedicated sandbox and spend-capped API key; it runs real shell commands and thinks around the clock.

Headlong is research software, be careful!

curl -fsSL https://t.co/P2Q5Ch8C4e | bash

Launch post: https://t.co/HhWULUTkpH

Repo: https://t.co/U4JunpSTYt

Headlong is a @LaudeInstitute / MIT collaboration.

[Andy Konwinski @andykonwinski]

I'll give $1M to the first open source AI that gets 90% on this sweet new contamination-free version of SWE-bench - https://t.co/o7LuYKIfhO

[Andy Konwinski @andykonwinski]

In the first few months Perplexity had a new product demo every 2-3 weeks. Product-market-fit in 4 months. The velocity permanently defined the culture.

[John-Michael Murphy @johnmichael_mur]

friday was my last day at @nytimes. it was a dream job. i'm so grateful for my colleagues and will miss them dearly.

tomorrow is my first day at @perplexity_ai. couldn't be more excited!

[Perplexity @perplexity_ai]

Perplexity Pro users can now experience @playground_ai's new V2.5 as the default model to generate images. Whether you need photographs, paintings, illustrations, diagrams, or something more custom to visually enhance your answer, the choice is yours!

[Perplexity @perplexity_ai]

Introducing pplx-api, our LLM API which serves Mistral and Llama2 models with blazing speed and throughput. pplx-api is in public beta for our Pro subscribers!

We partnered with @nvidia and @awscloud to build our proprietary inference. Learn more: https://t.co/fSII2O4QqU

[Perplexity @perplexity_ai]

Today, we're excited to introduce Collections:
a simple and intuitive way to organize your threads on Perplexity. 🔗 https://t.co/yQaQ6uYe1W

Collections lets you categorize threads and share them with other users who can contribute and browse. We're excited about this transition from being a tool to a knowledge platform with multi-player experiences.
Start creating and sharing!

[Perplexity @perplexity_ai]

Exciting update to Copilot on Perplexity with @OpenAI's GPT-3.5 fine-tuning API! We've improved speed, reduced cost, and matched GPT-4 performance. Copilot now engages faster without losing quality. Here's what you need to know: 🧵

[Perplexity @perplexity_ai]

Code LLaMA is now on Perplexity’s LLaMa Chat!

Try asking it to write a function for you, or explain a code snippet: 🔗 https://t.co/rwcPzknBgE

This is the fastest way to try @MetaAI’s latest code-specialized LLM. With our model deployment expertise, we are able to provide you with this model less than 24 hours of it’s release.

What’s next?
We’ll integrate code LLaMA into Perplexity, all in service of providing you with the best answers to your most technical questions!

[Perplexity @perplexity_ai]

We celebrate those committed to learning and growing smarter. As the new school year begins, we're excited to offer tips to help students and educators streamline their workflow.

[Perplexity @perplexity_ai]

Knowledge is power. Experience our new brand, designed to unlock our curiosity, one question at a time. Discover knowledge and explore endless possibilities: https://t.co/I6UdpJ2LBN

[Perplexity @perplexity_ai]

👩‍🔬 Now in Beta: File Upload

Try it now: https://t.co/wb1uut73JU

Start asking your questions directly to your documents, code, or research papers. Dive into information faster and get insights quicker. It's in beta, and we'd love your feedback.

[Perplexity @perplexity_ai]

🚨 LLaMa-2-70B-Chat is available now!

Try it here: https://t.co/2yYQB9jln7

💬 What will you ask first?

[Nick Jalbert @NickAJalbert]

RT @perplexity_ai: Introducing LLaMa Chat: The fastest way to chat with @MetaAI/@meta's Llama2!

Try it here: https://t.co/pVyd70OGxT, fee…

[Perplexity @perplexity_ai]

✨Now you can use Perplexity straight from your Chrome search bar! Install the Perplexity Default Search Chrome Plugin, and make every search a Perplexity query. Get the answer to (almost) every question, right away.

Get it now! ⬇️
https://t.co/UTrjJmQAbg

[Perplexity @perplexity_ai]

⚡️Supercharge your search with Perplexity Pro! Unlock nearly unlimited access to our most advanced search Copilot, the latest AI models, Pro support, and more - all for just $20/month or $200/year. 🚀
Get started at https://t.co/SA5jOdsP7J

[Perplexity @perplexity_ai]

Introducing your AI Profile on Perplexity! Personalize your AI answers with your bio, language, location, you name it. With dynamically generated questions, we're aiming to make your AI experience truly tailored to you. Dive in and experience Perplexity in a new personalized way today. https://t.co/9E3ICfgXS4

[Perplexity @perplexity_ai]

Perplexity for Android is here! 🎉 A superpower on your phone that allows you to search, discover, research and learn faster than ever before. Ask anything with voice input, learn with follow-up questions, and build your library of threads. Try it now at https://t.co/5GddQn83ej