Drew Schuyler

Dad of the best Girls. Shelly’s Husband. Pay attention. Be present. Be Kind

[Drew Schuyler @drewsky1]

https://t.co/LYC9HrAEfc

[Drew Schuyler @drewsky1]

@davideciffa Any idea on pricing?

[Drew Schuyler @drewsky1]

@grok @elonmusk I’d just like enough bandwidth to make him the orchestrator in @NousResearch Hermes

[Drew Schuyler @drewsky1]

Actually, after thinking about it for a bit, this is probably better as a plug-in. I have closed the PR.

[Drew Schuyler @drewsky1]

Opened a PR to Hermes Agent for receipt-backed candidate evaluation.
The lesson I've learned: don’t just test a model—test the model plus the inference stack it actually runs on, and keep receipts you can rescore offline.
Open for review, not merged:
https://t.co/4cP2arCf9W @Teknium - I think this is now way more robust

[Drew Schuyler @drewsky1]

@BrianRoemmele @NousResearch @Teknium - any interest?

[Drew Schuyler @drewsky1]

@BrianRoemmele maybe you should work with @NousResearch with that gigantic corpus of Pre Internet data you have. An open source model that is designed fully to run inside Hermes would be amazing!

[Drew Schuyler @drewsky1]

https://t.co/sMgxryNHfT this is the tip of the iceberg

[Brian Roemmele @BrianRoemmele]

I JUST GOT GIFTED 100s of 1000s OF MOSTLY UNPUBLISHED PHD DISSERTATION PAPERS ON MICROFICHE!

This is an exciting discovery!

In a metal basement cabinet of a former Ivy League physics professor was discovered shelves of microfiche at the great grandkids had no idea what to do.

On May 15, 2026 It was time for the multigenerational family vacation property to be cleaned and remodeled for August retreat. It was owned by the retired professor who passed away 12 years ago with his wife passing in 2023.

One of the grandkids who is now in his 40s was always curious by this cabinet so they got a locksmith to open it, perhaps the first time in 15 years and what it contained they didn’t quite understand.

That grandkid turned out to be somebody that is a https://t.co/tcKeuiQyql subscriber and he immediately identified he was looking at microfiche and something to do with scientific papers.

We had a conversation a few weeks back and my research has produced something that is absolutely blowing me away.

His grandpa, who, for many reasons I cannot name was conducting a wide ranging study of PhD dissertation papers from up to 16 universities. He ran a microfiche archiving operation for his work for decades.

I discovered he had private funding to conduct this research and it continued on well into his retirement. Unfortunately, a fire at his residence in Boston destroyed most of his manuscripts on what he was trying to achieve and the index of everything that was in a cabinet. It was his hope when he moved to their vacation home to to live at his days to restart his work unfortunately, it’s not clear if he ever did and there’s no evidence that anybody can find.

We can include that he discovered thousands of heavy gems and dissertation papers that never came to life, especially in the pre-Internet days.

The earliest entry I see is from 1873 all the way up to the early 2000s. All the papers are signed based and mostly in physics. The largest cohort is from the 1950s.

Some of the dissertations are handwritten, and most are manual typewritten. Some appear to be type set as a published book.

Most of them have not been digitalized or even saved. And this was one of the problems. The professor had is dissertations at many universities were discarded after a period of time in batches. And with it all of the work of the best students those universities had it was a space, savings equation, and belief that the old ideas are always replaced by the new ideas.

I’ve been collecting dissertation archives for decades, and this is the largest one I’ve come across. It is shocking to most people that most dissertations ever made prior to the year 2003 have not been digitalized and are mostly not available online.

I have permission to digitalize the entire archive for AI training only. I’m working with The Family to try to open source that data so that is freely available for everybody.

For obvious reasons, The Family wants to consult their attorney about open sourcing the data, but I’m already starting the digitalization process today.

We are building an AI that thinks unlike any other AI platform ever created because of the thought process that went into these dissertation papers. For it is the thought process more importantly than the actual data contained that we’re trying to capture in this AI platform. Although the data contained is unavailable anywhere else.

What is really heartbreaking is when I look at the data and almost every case it is the last copy of this PhD thesis that exists.

And if the stars did not align, and God did not guide that cabinet was scheduled to be taken away by a “Got Junk” truck most surely gone forever.

It is my hope that one day I get permission to name parts of this model after this professor.

More soon!

[Drew Schuyler @drewsky1]

@BrianRoemmele @elonmusk any chance you could use this data to create a good open source model for those of us in the US?

[Drew Schuyler @drewsky1]

@BrianRoemmele When are you going to train your own model on this? Are you thinking like a 9B or 27B or a MOE?

[Drew Schuyler @drewsky1]

New LuceBox PR #515: opt-in F32 rollback checkpoints for Qwen3.6-27B Q4 on one RTX 3090.

Fresh standalone test: +6.46% decode throughput. Defaults unchanged; experimental; ~1.15 GiB extra VRAM. I’m going to try to squeeze every bit out of these 3090s

@luceboxai @davideciffa @pupposandro #LocalAI
https://t.co/2Qt8XPU4pc

[Drew Schuyler @drewsky1]

I just submitted a DFLASH optimization for Qwen3.6-27B running locally on 2× RTX 3090s.

In controlled A/B/A testing:
• Decode speed: 22.76 → 29.41 tokens/sec (+29.2%)
 • Prompt processing: ~95.4 tokens/sec
 • End-to-end wall time: 9.4% lower
 • Output: byte-identical across all measured runs

The production patch adds a fast speculative-decoding rollback path using F32 SSM checkpoints. It passed the full CUDA build, all 15 current upstream tests, and targeted dual-GPU validation with 43/43 fast rollbacks succeeding and zero fallbacks.

Trade-off: approximately 1.65 GiB of additional persistent GPU memory. Both GPUs are required for this configuration; it will OOM on a single 24 GB RTX 3090.

The PR is now open for review—not merged yet:
https://t.co/WZtR6wdIJf

@luceboxai @davideciffa @pupposandro

[Drew Schuyler @drewsky1]

They keep promising abundance. They never promise it to you.

Every well-connected voice this year is selling the same future. Abundance everywhere. Scarcity solved. The wave lifts everyone. The trajectory is real. The models are real. The acceleration is real.

The “everyone” is the lie, and the lie is moving faster than anyone selling it will admit.

This week the administration asked OpenAI to release its best model customer by customer — approving access one institution at a time — and OpenAI agreed. Not one user at a time. One customer. You are not a customer. Per The Information’s reporting, the request came over security concerns from inside the federal review process. Same stretch: Anthropic’s most capable models, pulled from the public entirely. The frontier didn’t slow down. It walked through a door marked enterprise and government only, and the door locked behind it.

Cheap intelligence is not cheap grain. You cannot ration bread at the source. You can ration intelligence with a license, a login, a customer list. Abundance you need permission to reach is not abundance. It’s a tap, and someone else owns the valve.

Today there’s a way around the valve. Open weights. A model you download, run on hardware you own, and use without anyone approving you first. This week an open model matching last year’s frontier shipped free and MIT-licensed, small enough to run in a basement. That is the only abundance that’s actually abundant — the kind nobody can meter.

So look at what gets gated next.

The capability that triggers the government review is cyber capability. Every model good enough to be the next real step is, by construction, good enough at code and exploitation to trip that wire. The gate that’s on closed models today lands on the open ones the moment they cross the same line. And the tool for that already exists. Weights can be export-controlled. Hosting them can be restricted. There’s a bill that already passed the House treating remote access to controlled compute as an export. The machinery to ban an open model from American hands is built. It’s waiting for a model worth banning.

When that day comes, they will not call it a ban. They will call it security. They have run this exact play before. Encryption sat on the federal munitions list next to bombs and flamethrowers, and a grad student who wanted to publish his own code was told he was an arms dealer. He sued in the open. He won. Code is speech. That ruling is the reason your bank login works. The pattern is old and it is consistent: power meets a technology that redistributes power, calls it a weapon, and meters it — until someone forces it open in daylight.

That fight is coming for open weights, and it will come fast, because everything is coming fast now and no one in a position to slow it down has any incentive to.

The abundance is real. The question was never whether. It’s who holds a key when the doors close — and the answer is the people who already hold the most, unless the rest of us own the thing outright before they decide we’re not allowed to.

One path needs permission. The other needs a power cable. Get the second one running while it’s still possible.

[Drew Schuyler @drewsky1]

@alexwg ,@PeterDiamandis is the underlying idea worthy of a Moonshots discussion?

[Drew Schuyler @drewsky1]

https://t.co/PBJLw1bXJ4

[Drew Schuyler @drewsky1]

@alexwg Who is worthy to accept this quest and travel to Mordor to return this model to whence it came…

[Drew Schuyler @drewsky1]

@vibeeval @g_kassap iOS when?

[Drew Schuyler @drewsky1]

@alexwg the asteroid belt will become the asset,once we flywheel autonomus mining/manufacturing. the product can be mass driven to it's respective place in the Dyson Swarm