Jason Gauci
@NeuralNets4Life
Followers
1K
Following
501
Media
174
Statuses
724
I bring AI to billions. Reinforcement Learning @ Meta AI, Google AI. Programming Throwdown host. Eternal Terminal, MAMEHub, ReAgent creator
Austin, TX
Joined April 2010
I'm playing Symphony of War. The thing about indie tactical RPG games is that you never know when the difficulty curve is going to ramp up unexpectedly. But it's also kind of thrilling. Is the game beatable or have poor choices early on soft-locked it? Only one way to find out!.
0
0
2
LinkedIn is Oxford dictionary. X is urban dictionary. Don't believe me? Ask the anime chatbot.
0
0
2
It was a blast speaking with @jake_oshea about Reinforcement learning, self-driving cars, and of course Circuit!.
We are LIVE! Excited to share that the first episode of the @aitxcommunity podcast is now available online!. Our first episode is with @NeuralNets4Life, an OG AI researcher and engineer. In this conversation, we discuss Jason's personal background and entrance into the Machine.
1
0
3
Going hiking in the Bay Area:. 1. Pack everything into a car.2. Drive up the hills. Tech bro behind you wants to show off their Lambo. One lane road. Horns honking. 3. Fight for a parking spot. Road rage. 4. Hike. Going hiking in Austin. 1. Jump the backyard fence. 2. Hike
2
0
2
The lady at the whole food checkout counter asked how my day was. It's like she wanted Chat to just respond right away or something. So impatient, get some good wifi Whole Foods.
1
0
3
I went back and listened to episode 183 on getting a job in tech. Easily the best soft-skills episode we've ever done. 100 minutes of truth nukes (76.9 min at 1.3x). Send it to anyone who is looking for work! Infinitely worth the price.
0
0
3
In my 20s, one job: AI researcher. It went really well. In our 40s, many jobs: home repair, lucha libre fighter, gardener, maintainer of open source, math tutor. it's not dilution: it's alchemy. New jobs give us perspective. We avoid bad teams and make our teams unavoidable.
1
0
13
I am super excited to have Mark Cunningham ( @mhcrecruit ) on the show to talk about how to land a software engineering job in 2025! The three of us dive into the four stages of interviewing, offering a bunch of tips and advice!. Show notes:
0
0
1
Bricks & Minifigs Bee Cave is hosting a #LEGOMastersFOX watch party! Good luck Ian and Sage! Go #austin!
0
0
0
Reading through the papers at this year's @RL_Conference, and found this one on an exploration platform to be really interesting! Check it out:.
arxiv.org
Extrinsic rewards can effectively guide reinforcement learning (RL) agents in specific tasks. However, extrinsic rewards frequently fall short in complex environments due to the significant human...
0
0
1
I'm so excited for the episode we recorded on interviewing (with an actual recruiter guest host) to go live!!! As soon as post-production is done, I'm pressing the big green button. Let's go!
1
0
2
Reinforcement learning wasn't originally made to work with neural networks. DQN/PPO was, but are hard to train. Neural nets are biased & the critic model adds bias, making it worse. GRPO reduces variance without a critic model by critiquing each training example on the fly.
0
0
3
RT @rockets4all2: @Austen Scrape GitHub for commits then have ai guess how complicated these commits are. Pick the top 2% according to ai’s….
0
1
0
Patrick and I have been BLOWN AWAY by the feedback from the last intro where we talked about how to land a job. This is definitely an area that needs more attention, so on Friday we invited one of my favorite recruiters to come on the show and chat! What should we ask him?
0
0
1
😂.
@xai We’re all looking for the guy who did this, be on the lookout for someone who: . 1. Has admin access to key functionality.2. Seems to disregard authority/process.3. Possibly some connection to S. Africa.
0
0
1
RT @war24182236: 🎥 July 6 2025 San Francisco infamous .Tenderloin on Leavenworth between Eddy and Ellis Street across the street from the T….
0
87
0
What are some good, recent papers in #reinforcementlearning that aren't about LLM post-training?.
0
0
1
Great talk about where reinforcement learning has been and where it will go in the future.
In my recent talk at the Upperbound conference, I included a slide in which I tried to be a realist about the arrival of AI, setting aside what we might want to happen or what we might fear will happen, and just ask what _will_ happen (as in John Mearsheimer's "realist" school of
0
0
1
RT @SahilBloom: This is one of my favorite movie scenes of all time. It always hits me hard:. "No I didn't rush the f*cking field, I was….
0
569
0