Pinned Tweet
Lascari Live, Wednesday @ 6pm CT
Build a company brain with one of Austin’s best AI FDEs
Ford Lascari will lead a hands-on workshop on building agent systems around real company context, tools, and workflows
Apply below
🍄DB
three weeks ago this project didn't exist.
today mushroomdb is an embedded graph database written in rust built for apps and agents that need to keep track of how things relate.
self maintaining edges, time travel queries, per caller access control, an explain function that shows why any two nodes are linked, etc
it lives inside your app like sqlite, cold opens in 20ms, and you can install it today with cargo, npm or pip.
the general idea for me was to stop handwriting relationships. declare the connections that matter once and the engine creates, updates, and retracts them as your data changes all deterministically
ask it why any edge exists and it shows the arithmetic. some of this i genuinely haven't seen other databases do
the tooling available now is the reason this took three weeks instead of six months. building in 2026 is a completely different ballgame than building in 2024.
still pre-alpha state but the fun an interesting parts already work and it's only getting better from here everyday
link to the repo in reply.
Our first Hack AI Workshop is this Wednesday in Austin.
Come with a real business problem. Leave with something working.
Applications below ↓
I'm excited to kick off our first Hack AI Workshop.
After two years of bringing the best AI talent in Austin together, we're rolling out smaller, hands-on events focused on actually building.
The goal is simple: bring a real business problem and leave with something working.
Linkt AI FDEs will be in the room to share best practices, then work alongside attendees as they build.
Our first workshop is this Wednesday, led by @Ford_Lascari, where attendees will build a "company brain" in under two hours.
Apply here: luma.com/wpprtdtt
Linkt retweeted
Be there! 🤠
We are excited to host Hack AI’s AI Startup Gathering this Wednesday, March 11th from 6pm-8pm!
The event is sponsored by @targoncompute @_linktai and others, attendees will connect and learn about accelerating Austin’s AI Space!
We look forward to seeing you soon!
Linkt retweeted
TLDR
Clients -> Partners
Seat based Subscription -> A Stake in the Outcome
Talk soon @garysheng
This Post is from an account that no longer exists. Learn more
👀
OpenBioLLM-8B and OpenBioLLM-70B are new fine-tuned Llama models developed by Saama to streamline tasks that can accelerate clinical trials, opening up new possibilities in personalized medicine. More details in their research paper ➡️ go.fb.me/x3iq84
Linkt retweeted
Our team @_linktai pushed more code than ever this year...despite only growing headcount by 2 developers
Our stack:
- Cursor (w/ sonnet)
- Granola: call transcript to Jira ticket FAST
- ChatGPT Pro Mode
Cant even imagine what 2025 is going to look like
Sonus AI has launched the Sonus-1 series of large language models (LLMs), including Sonus-1 Mini for speed, Sonus-1 Air for balanced performance, and the high-performance Sonus-1 Pro with advanced reasoning capabilities.
The Sonus-1 Pro achieved 87.5% on the MMLU benchmark, rising to 90.15% with reasoning. It scored 71.8% on MMLU-Pro and excelled in math with 91.8% on MATH-500. Additionally, it achieved 88.9% on DROP and 67.3% on GPQA-Diamond, showcasing its competitive edge.
These advancements highlight Sonus AI's commitment to delivering high-performance models for diverse applications. For more details, visit sonus.ai/blog/sonus-1
