Samad Syed (@SamadSyed) header
Samad Syed (@SamadSyed) profile picture

@WisconsinCS ‘26 / Prev AWS / UW Research Train in systems @getcrackedio

Joined December 2025
Photo by Samad Syed (@SamadSyed)Photo by Samad Syed (@SamadSyed)Photo by Samad Syed (@SamadSyed)Photo by Samad Syed (@SamadSyed)Photo by Samad Syed (@SamadSyed)Photo by Samad Syed (@SamadSyed)Photo by Samad Syed (@SamadSyed)Photo by Samad Syed (@SamadSyed)Photo by Samad Syed (@SamadSyed)Photo by Samad Syed (@SamadSyed)
StudyLoop is going to NeurIPS :) Accepted to two workshops: • Verify-Agents • RAAAI Single author. Independent research. Now I just gotta figure out how to get to Sydney and Atlanta 😭
Can an AI agent study a codebase before it knows the questions it’ll be asked? And if so, which techniques work best today? I built StudyLoop to test that. The results were interesting. 🧵
1
4
2,775
Perplexity to 1T by mid 2028
19
Samad Syed retweeted
Will stronger LLMs make harnesses obsolete? New MIT CSAIL paper argues the opposite: Harnesses compose model capabilities to generalize across longer, deeper, & more strategic tasks, while enabling more data-efficient training: tinyurl.com/2fwk4uap
14
37
5
261
12,989
Beautiful
One isolated Postgres branch per pull request, ready in under a minute. See how @SnorkelAI uses Lakebase copy-on-write branching for CI/CD testing, designed to support around 100 concurrent PR branches without duplicating the full dataset for each one. databricks.com/blog/lakebase…
1
149
Loved these two graphics from @DbrxMosaicAI MemCurrent blog, I think derived memory is the future and should be the focus.
42
Dave Patterson has a habit of giving bad advice. His latest, How to Earn a Bad Obituary, explains why he's joining @LaudeInstitute as a Member of Technical Staff: to work for the public good with all the urgency of a startup. First on his to-do list: leading our next batch of Moonshots! 🌕 More good (bad) advice from Dave: laude.org/updates/how-to-ear…
2
99
Samad Syed retweeted
Rashid Bashir, dean of the Grainger College of Engineering and vice chancellor for Chicago strategic partnerships at the University of Illinois-Urbana Champaign, has been named as the next chancellor of the University of Wisconsin–Madison.   news.wisc.edu/rashid-bashir-…
32
42
15
250
94,277
So excited!
I've never in my life been busier than the past several weeks, and I apologize if I've been slow with anyone, but the various remaining threads are all converging - and I think y'all will love them.
1
154
Samad Syed retweeted
We’ve heard again and again that enterprises don’t want to pour all their valuable IP, data and tacit knowledge into someone else’s model. They want control of their own weights.
118
130
73
2,907
363,735
Samad Syed retweeted
Introducing Qwen3.8-27B-pi-adapted 🤗 We adapted Qwen's model specifically for the pi coding agent using reinforcement learning and a LoRA adapter. The results, measured as pass@1: 🚀 SWE-bench Pro (Hard 51): 29.4% → 62.7% 💻 Terminal-Bench 2.1: 39.5% → 55.8% ⚡ 55–62% fewer tool calls per task This is the power of open models: you can adapt them to YOUR tools and YOUR use cases. Let’s build! 💛 Open weights: huggingface.co/harness-hub/Q… Methodology: huggingface.co/blog/harness-…
308
643
175
5,998
456,846
I remember reading this back when I was in undergrad. A really good recipe.
hot take: nothing wrong with papers per se, but… here’s the recipe github.com/okhat/blog/blob/m…
1
8
99
11,661
oasys lab is shipping research at higher velocity than 99.9% of startups ship product nice work @lateinteraction - super inspiring. also my vote for best name on the list
most AGI pilled labs: - typesafe - general reasoning - general intuition - mit oasys - amp code
1
7
44
5,753
terminal bench and TB science voted #1 and #2 for "benchmarks trusted most to indicate if an LLM is frontier"
I asked 56 people in AI which benchmarks they'd trust most to indicate if an LLM is frontier: 25 votes: Terminal-Bench 4 (@alexgshaw, @andykonwinski, @Mike_A_Merrill) 20 votes: Terminal-Bench Science (@StevenDillmann, @lschmidt3, @sanmikoyejo) 13 votes: FrontierCode (@penlume, @ybenpan, @denizbirlikci, @silasalberti) 10 votes: FrontierSWE (@evan_j_chu, @_rajanagarwal, @MatternJustus, @ardnoKhtnayhsiR, @sanket__mhatre, @akshitkr) 8 votes: Humanity's Last Exam (@longphann_, @hendrycks, Alice Gatti, @ziwen_h, @natliml) 7 votes: Vending-Bench (@axelbacklund, @lukaspet) 5 votes each: - AutomationBench (@danielwshepard, @SalimansRobin, @TweetAnnaMarie) - DeepSWE (@winkey_h, @charleyslee, @LeonardMainnet, @serenaa_ge) - PostTrainBench (@full__rank, @hrdkbhatnagar, @AmyPrb, @shiraeis, @karinanguyen, @MatthiasBethge, @maksym_andr) The most common comment was that people don't really trust benchmarks at all. 56 people voted. People couldn't pick a bench their company worked on, an index, or a human-voted board. Disclosure: on this list, I'm a small investor in Abundant, Cognition, Emulated, Fleet, Hillclimb, Latch, and Snorkel. I didn't vote. Counts are net: a few benchmarks received negative votes. Continued:
2
7
1
62
5,028
Wow, this post really took off. Thank you everyone for the support! For my new followers, I'm an independent AI researcher. I mostly tweet about technical topics in open-source AI, my own work, interesting work from others, and the occasional vaguepost. Hope you stick around!
Unfortunately, as an independent researcher, I can't afford to attend NeurIPS 2026 in Sydney or Atlanta to present my accepted papers. I applied for several travel grants and financial aid opportunities, but none worked out. If anyone attending either event would be willing to put up my poster, I'd be incredibly grateful. Please reach out!
2
278
Unfortunately, as an independent researcher, I can't afford to attend NeurIPS 2026 in Sydney or Atlanta to present my accepted papers. I applied for several travel grants and financial aid opportunities, but none worked out. If anyone attending either event would be willing to put up my poster, I'd be incredibly grateful. Please reach out!
14
7
3
229
38,673
It pays off to be good at C++, low level systems, and high performance code
Replying to @elonmusk @tetsuoai
At SpaceX, we write our own headers and calculate our own time (includes leap seconds). Not many people will know what this means, but those that do will appreciate it. The new Grok AI/SI training and inference stack is derived from incredible of work done by amazing engineers to create Starlink.
3
555
Samad Syed retweeted
Fun Fact: Basically none of the initial railroad companies made money. They overbuilt, took on way too much debt, and ended up in brutal pricing competition with each other, and nearly all went bankrupt.
201
439
100
10,813
531,684
remzi understands that as the AI paradigm advances, the decisions about what we build, and who we’re building for, will impact billions of lives. he was my undergrad advisor at UW-Madison. i met him and andrea when I took their operating systems class. i remember during section how he was a total master of the terminal (tcsh) and emacs and i promised myself i’d match his skillset some day. i started turning my mouse upside down in the computer labs to learn keyboard commands faster. after all these years, it’s an honor that i still get to work with @RArpaciDusseau, and also to have him advise @LaudeInstitute, where our open research residents get to benefit from his perspective just like i have.
“The root of academic freedom is to be able to look into the world with clear eyes and to speak the truth. If universities can’t do it, then who will?” Laude Research Advisor Remzi Arpaci-Dusseau, a longtime @UWMadison professor, is the founding Dean of @uwmadcai. At a key moment for computer science, AI, and the world, he sees universities as well-positioned to challenge ideas and ensure technology serves the public good. Worth a read: onwisconsin.uwalumni.com/the…
2
2
29
2,004