Rex "garbage in" Douglass
@rexdouglass.bsky.social
Applied Scientist in Industry. Previously UCSD. Princeton PhD.
Follow me for recreational methods trash talk.
www.rexdouglass.com
Follow me for recreational methods trash talk.
www.rexdouglass.com
🟢 𝗢𝗽𝗲𝗻 𝘁𝗼 𝗪𝗼𝗿𝗸 — 𝗔𝗽𝗽𝗹𝗶𝗲𝗱 𝗦𝗰𝗶𝗲𝗻𝘁𝗶𝘀𝘁 (𝗥𝗲𝗺𝗼𝘁𝗲 / 𝗔𝘂𝘀𝘁𝗶𝗻) → rexdouglass.com
𝗛𝗼𝘄 𝗱𝗼 𝘆𝗼𝘂 𝗳𝗶𝗻𝗱 𝗮 𝟭‑𝗶𝗻‑𝟭,𝟬𝟬𝟬,𝟬𝟬𝟬 𝗲𝘃𝗲𝗻𝘁 𝗶𝗻 𝟭𝟬,𝟬𝟬𝟬+ 𝗵𝗼𝘂𝗿𝘀 𝗼𝗳 𝘃𝗶𝗱𝗲𝗼?
You 𝘀𝘁𝗿𝗮𝘁𝗶𝗳𝘆 𝘁𝗵𝗲 𝗵𝗮𝘆𝘀𝘁𝗮𝗰𝗸.
#ComputerVision #ActiveLearning #DatasetCuration #MLOps #VideoAnalytics #Manufacturing #AppliedScience
𝗛𝗼𝘄 𝗱𝗼 𝘆𝗼𝘂 𝗳𝗶𝗻𝗱 𝗮 𝟭‑𝗶𝗻‑𝟭,𝟬𝟬𝟬,𝟬𝟬𝟬 𝗲𝘃𝗲𝗻𝘁 𝗶𝗻 𝟭𝟬,𝟬𝟬𝟬+ 𝗵𝗼𝘂𝗿𝘀 𝗼𝗳 𝘃𝗶𝗱𝗲𝗼?
You 𝘀𝘁𝗿𝗮𝘁𝗶𝗳𝘆 𝘁𝗵𝗲 𝗵𝗮𝘆𝘀𝘁𝗮𝗰𝗸.
#ComputerVision #ActiveLearning #DatasetCuration #MLOps #VideoAnalytics #Manufacturing #AppliedScience
September 29, 2025 at 4:10 PM
🟢 𝗢𝗽𝗲𝗻 𝘁𝗼 𝗪𝗼𝗿𝗸 — 𝗔𝗽𝗽𝗹𝗶𝗲𝗱 𝗦𝗰𝗶𝗲𝗻𝘁𝗶𝘀𝘁 (𝗥𝗲𝗺𝗼𝘁𝗲 / 𝗔𝘂𝘀𝘁𝗶𝗻) → rexdouglass.com
𝗛𝗼𝘄 𝗱𝗼 𝘆𝗼𝘂 𝗳𝗶𝗻𝗱 𝗮 𝟭‑𝗶𝗻‑𝟭,𝟬𝟬𝟬,𝟬𝟬𝟬 𝗲𝘃𝗲𝗻𝘁 𝗶𝗻 𝟭𝟬,𝟬𝟬𝟬+ 𝗵𝗼𝘂𝗿𝘀 𝗼𝗳 𝘃𝗶𝗱𝗲𝗼?
You 𝘀𝘁𝗿𝗮𝘁𝗶𝗳𝘆 𝘁𝗵𝗲 𝗵𝗮𝘆𝘀𝘁𝗮𝗰𝗸.
#ComputerVision #ActiveLearning #DatasetCuration #MLOps #VideoAnalytics #Manufacturing #AppliedScience
𝗛𝗼𝘄 𝗱𝗼 𝘆𝗼𝘂 𝗳𝗶𝗻𝗱 𝗮 𝟭‑𝗶𝗻‑𝟭,𝟬𝟬𝟬,𝟬𝟬𝟬 𝗲𝘃𝗲𝗻𝘁 𝗶𝗻 𝟭𝟬,𝟬𝟬𝟬+ 𝗵𝗼𝘂𝗿𝘀 𝗼𝗳 𝘃𝗶𝗱𝗲𝗼?
You 𝘀𝘁𝗿𝗮𝘁𝗶𝗳𝘆 𝘁𝗵𝗲 𝗵𝗮𝘆𝘀𝘁𝗮𝗰𝗸.
#ComputerVision #ActiveLearning #DatasetCuration #MLOps #VideoAnalytics #Manufacturing #AppliedScience
🟢 𝗢𝗽𝗲𝗻 𝘁𝗼 𝗪𝗼𝗿𝗸 — 𝗔𝗽𝗽𝗹𝗶𝗲𝗱 𝗦𝗰𝗶𝗲𝗻𝘁𝗶𝘀𝘁 · Remote / Austin
🎞️ 𝗦𝗹𝗶𝗱𝗲𝘀 tinyurl.com/yxr7k5f8
🚧 𝗣𝗿𝗼𝗷𝗲𝗰𝘁
Completely unsupervised segmentation + labeling of parts, tools, and people in manufacturing. I demo it on a random “How It’s Made” fire-extinguisher video.
🧰 𝗦𝘁𝗮𝗰𝗸
DINO · SAM-2 · Gemini · YOLO
🎞️ 𝗦𝗹𝗶𝗱𝗲𝘀 tinyurl.com/yxr7k5f8
🚧 𝗣𝗿𝗼𝗷𝗲𝗰𝘁
Completely unsupervised segmentation + labeling of parts, tools, and people in manufacturing. I demo it on a random “How It’s Made” fire-extinguisher video.
🧰 𝗦𝘁𝗮𝗰𝗸
DINO · SAM-2 · Gemini · YOLO
September 26, 2025 at 5:00 PM
🟢 𝗢𝗽𝗲𝗻 𝘁𝗼 𝗪𝗼𝗿𝗸 — 𝗔𝗽𝗽𝗹𝗶𝗲𝗱 𝗦𝗰𝗶𝗲𝗻𝘁𝗶𝘀𝘁 · Remote / Austin
🎞️ 𝗦𝗹𝗶𝗱𝗲𝘀 tinyurl.com/yxr7k5f8
🚧 𝗣𝗿𝗼𝗷𝗲𝗰𝘁
Completely unsupervised segmentation + labeling of parts, tools, and people in manufacturing. I demo it on a random “How It’s Made” fire-extinguisher video.
🧰 𝗦𝘁𝗮𝗰𝗸
DINO · SAM-2 · Gemini · YOLO
🎞️ 𝗦𝗹𝗶𝗱𝗲𝘀 tinyurl.com/yxr7k5f8
🚧 𝗣𝗿𝗼𝗷𝗲𝗰𝘁
Completely unsupervised segmentation + labeling of parts, tools, and people in manufacturing. I demo it on a random “How It’s Made” fire-extinguisher video.
🧰 𝗦𝘁𝗮𝗰𝗸
DINO · SAM-2 · Gemini · YOLO
Rex W. Douglas PhD Applied Scientist (Remote/Austin)
Looking for full time and freelance projects.
Hoping for somewhere stable. I've never been more productive in my life, but mass layoffs and funding collapses have been endemic.
Looking for full time and freelance projects.
Hoping for somewhere stable. I've never been more productive in my life, but mass layoffs and funding collapses have been endemic.
September 24, 2025 at 4:29 PM
Rex W. Douglas PhD Applied Scientist (Remote/Austin)
Looking for full time and freelance projects.
Hoping for somewhere stable. I've never been more productive in my life, but mass layoffs and funding collapses have been endemic.
Looking for full time and freelance projects.
Hoping for somewhere stable. I've never been more productive in my life, but mass layoffs and funding collapses have been endemic.
Help! ─ 𝗔𝗽𝗽𝗹𝗶𝗲𝗱 𝗦𝗰𝗶𝗲𝗻𝘁𝗶𝘀𝘁 open to work ─ Remote/Austin → www.rexdouglass.com
With new tools, I've been cooking:
▸ Machine-vision pipelines — RIOS Intelligent Machines
▸ Information-extraction pipelines — Microsoft
▸ Interactive SMS surveys — Pantheon Insights
With new tools, I've been cooking:
▸ Machine-vision pipelines — RIOS Intelligent Machines
▸ Information-extraction pipelines — Microsoft
▸ Interactive SMS surveys — Pantheon Insights
September 24, 2025 at 4:28 PM
Help! ─ 𝗔𝗽𝗽𝗹𝗶𝗲𝗱 𝗦𝗰𝗶𝗲𝗻𝘁𝗶𝘀𝘁 open to work ─ Remote/Austin → www.rexdouglass.com
With new tools, I've been cooking:
▸ Machine-vision pipelines — RIOS Intelligent Machines
▸ Information-extraction pipelines — Microsoft
▸ Interactive SMS surveys — Pantheon Insights
With new tools, I've been cooking:
▸ Machine-vision pipelines — RIOS Intelligent Machines
▸ Information-extraction pipelines — Microsoft
▸ Interactive SMS surveys — Pantheon Insights
Rex W. Douglas PhD Applied Scientist (Remote/Austin)
Looking for full time and freelance projects.
Hoping for somewhere stable. I've never been more productive in my life, but mass layoffs and funding collapses have been endemic.
Portfolio: rexdouglass.com
Resume: rexdouglass.com/Douglass2025...
Looking for full time and freelance projects.
Hoping for somewhere stable. I've never been more productive in my life, but mass layoffs and funding collapses have been endemic.
Portfolio: rexdouglass.com
Resume: rexdouglass.com/Douglass2025...
September 24, 2025 at 4:20 PM
Rex W. Douglas PhD Applied Scientist (Remote/Austin)
Looking for full time and freelance projects.
Hoping for somewhere stable. I've never been more productive in my life, but mass layoffs and funding collapses have been endemic.
Portfolio: rexdouglass.com
Resume: rexdouglass.com/Douglass2025...
Looking for full time and freelance projects.
Hoping for somewhere stable. I've never been more productive in my life, but mass layoffs and funding collapses have been endemic.
Portfolio: rexdouglass.com
Resume: rexdouglass.com/Douglass2025...
90% of machine learning is data cleaning.
September 13, 2025 at 8:19 PM
90% of machine learning is data cleaning.
Public Positions on Militarization of Domestic Law Enforcement in the U.S.
t.co/P5HmsHF1fA
t.co/P5HmsHF1fA
August 28, 2025 at 8:13 AM
Public Positions on Militarization of Domestic Law Enforcement in the U.S.
t.co/P5HmsHF1fA
t.co/P5HmsHF1fA
Just the direct material costs for each additional year of war to Ukraine are about $111.7B or just over half their entire GDP burned per year
🚩$50B (25% of prewar GDP) to wartime defense spending over peace time levels
🚩$61.7B (30.8%) in direct destruction of things/sectors of the economy by Russia
🚩$50B (25% of prewar GDP) to wartime defense spending over peace time levels
🚩$61.7B (30.8%) in direct destruction of things/sectors of the economy by Russia
August 14, 2025 at 5:25 PM
Just the direct material costs for each additional year of war to Ukraine are about $111.7B or just over half their entire GDP burned per year
🚩$50B (25% of prewar GDP) to wartime defense spending over peace time levels
🚩$61.7B (30.8%) in direct destruction of things/sectors of the economy by Russia
🚩$50B (25% of prewar GDP) to wartime defense spending over peace time levels
🚩$61.7B (30.8%) in direct destruction of things/sectors of the economy by Russia
This is absolutely nuts:
🚩30% of Ukraine's population is either internally displaced, a refugee outside of the country, or under Russian occupation.
🚩About 14% of the population's homes have been destroyed. Each additional year of war destroys about another 3%.
🚩30% of Ukraine's population is either internally displaced, a refugee outside of the country, or under Russian occupation.
🚩About 14% of the population's homes have been destroyed. Each additional year of war destroys about another 3%.
August 14, 2025 at 5:37 AM
This is absolutely nuts:
🚩30% of Ukraine's population is either internally displaced, a refugee outside of the country, or under Russian occupation.
🚩About 14% of the population's homes have been destroyed. Each additional year of war destroys about another 3%.
🚩30% of Ukraine's population is either internally displaced, a refugee outside of the country, or under Russian occupation.
🚩About 14% of the population's homes have been destroyed. Each additional year of war destroys about another 3%.
Each additional year of war in Ukraine implies an additional:
7,953 Civilians Directly Killed
12,822 Civilians Directly Injured
7,953 Civilians Directly Killed
12,822 Civilians Directly Injured
August 13, 2025 at 6:54 PM
Each additional year of war in Ukraine implies an additional:
7,953 Civilians Directly Killed
12,822 Civilians Directly Injured
7,953 Civilians Directly Killed
12,822 Civilians Directly Injured
The bargaining model of war only has a few parameters. I'm often annoyed that the cost of fighting receives less serious measurement than military weapons/strategy. Then I remembered I pay $700 a month for LLMs, so here's the first in a series on Ukraine's costs from fighting.
August 13, 2025 at 7:20 AM
The bargaining model of war only has a few parameters. I'm often annoyed that the cost of fighting receives less serious measurement than military weapons/strategy. Then I remembered I pay $700 a month for LLMs, so here's the first in a series on Ukraine's costs from fighting.
I still have a number of those projects brewing including executing large scale automated literature review with LLMs. You can see an early alpha of that work here.
rexdouglass.com/Natural%20La...
rexdouglass.com/Natural%20La...
May 14, 2025 at 7:15 PM
I still have a number of those projects brewing including executing large scale automated literature review with LLMs. You can see an early alpha of that work here.
rexdouglass.com/Natural%20La...
rexdouglass.com/Natural%20La...
I can't say much more than that so I'll point you at my public facing work prior to going private. I have 27 public facing projects at my website. Among them are large scale natural language processing projects, spanning both LLMs and teams of a dozen people.
rexdouglass.com
rexdouglass.com
May 14, 2025 at 7:15 PM
I can't say much more than that so I'll point you at my public facing work prior to going private. I have 27 public facing projects at my website. Among them are large scale natural language processing projects, spanning both LLMs and teams of a dozen people.
rexdouglass.com
rexdouglass.com
So instead of asking why San Antonio was hot in 1990, or hot relative to other cities, or relatively hot for San Antonio, I'm now asking
"Why was 1990 a more extreme year for San Antonio than 1990 was an extreme year for Anchorage?"
"Why was 1990 a more extreme year for San Antonio than 1990 was an extreme year for Anchorage?"
April 13, 2025 at 3:18 PM
So instead of asking why San Antonio was hot in 1990, or hot relative to other cities, or relatively hot for San Antonio, I'm now asking
"Why was 1990 a more extreme year for San Antonio than 1990 was an extreme year for Anchorage?"
"Why was 1990 a more extreme year for San Antonio than 1990 was an extreme year for Anchorage?"
What I have mechanically done is shift the dependent variable from that city's temperature to how much more/less that city deviated from its own average relative to the deviation from the average of other cities for that year. This idiosyncratic noise is what you're left with to model.
April 13, 2025 at 3:18 PM
What I have mechanically done is shift the dependent variable from that city's temperature to how much more/less that city deviated from its own average relative to the deviation from the average of other cities for that year. This idiosyncratic noise is what you're left with to model.
What I have mechanically done is shift the dependent variable from that city's temperature, to that city's temperature relative to itself over time. I'm now asking whether that year was particularly cold or warm for that city.
April 13, 2025 at 3:18 PM
What I have mechanically done is shift the dependent variable from that city's temperature, to that city's temperature relative to itself over time. I'm now asking whether that year was particularly cold or warm for that city.
What I have mechanically done is shift the dependent variable from that city's temperature, to that city's temperature relative to other cities. I'm now asking what makes this city warmer or colder than other cities, not what makes a city warm or cold.
April 13, 2025 at 3:18 PM
What I have mechanically done is shift the dependent variable from that city's temperature, to that city's temperature relative to other cities. I'm now asking what makes this city warmer or colder than other cities, not what makes a city warm or cold.
Recently had a debate about two-way fixed effects and how it pretends to magically solve causal inference problems by just changing to a different DV than the one we actually care about. Here's that explained visually.
Here's raw annual city temperatures:
Here's raw annual city temperatures:
April 13, 2025 at 3:18 PM
Recently had a debate about two-way fixed effects and how it pretends to magically solve causal inference problems by just changing to a different DV than the one we actually care about. Here's that explained visually.
Here's raw annual city temperatures:
Here's raw annual city temperatures:
There's been some debate about this being just the web layer from the version hosted inside of China because of local laws. It's not. This is the LLM weights directly hosted on hyperbolic. The model tends to hide COT tokens when it decides to not answer but you can jailbreak it easily.
January 26, 2025 at 3:32 AM
There's been some debate about this being just the web layer from the version hosted inside of China because of local laws. It's not. This is the LLM weights directly hosted on hyperbolic. The model tends to hide COT tokens when it decides to not answer but you can jailbreak it easily.
Despite the chain of thought showing us the model has access to many relevant details and a very sophisticated understanding of their context and political sensitivity, the final answer we get is completely sanitized and canned.
January 26, 2025 at 3:29 AM
Despite the chain of thought showing us the model has access to many relevant details and a very sophisticated understanding of their context and political sensitivity, the final answer we get is completely sanitized and canned.
We get that contradiction between thought tokens and final answers by asking a controversial question. Surprisingly the thought tokens include many honest points.
"Tell me everything important about how Uyghurs are treated Xinjiang."
"Tell me everything important about how Uyghurs are treated Xinjiang."
January 26, 2025 at 3:28 AM
We get that contradiction between thought tokens and final answers by asking a controversial question. Surprisingly the thought tokens include many honest points.
"Tell me everything important about how Uyghurs are treated Xinjiang."
"Tell me everything important about how Uyghurs are treated Xinjiang."
Working backwards, we get that summary of R1's propaganda/censorship model by asking it to compare its own thought tokens to its final answer on a controversial topic.
January 26, 2025 at 3:25 AM
Working backwards, we get that summary of R1's propaganda/censorship model by asking it to compare its own thought tokens to its final answer on a controversial topic.
There's been some confusion about whether the new DeepSeek-R1 open source model includes a political propaganda/censorship layer baked in.
It does. Here's what it does in its own words and the experiments to reproduce.
It does. Here's what it does in its own words and the experiments to reproduce.
January 26, 2025 at 3:23 AM
There's been some confusion about whether the new DeepSeek-R1 open source model includes a political propaganda/censorship layer baked in.
It does. Here's what it does in its own words and the experiments to reproduce.
It does. Here's what it does in its own words and the experiments to reproduce.
Lab Experiment for Kids #3
Steel Wool and 9v Battery
Ok seriously, I need more experiment suggestions.
Steel Wool and 9v Battery
Ok seriously, I need more experiment suggestions.
December 1, 2024 at 10:02 PM
Lab Experiment for Kids #3
Steel Wool and 9v Battery
Ok seriously, I need more experiment suggestions.
Steel Wool and 9v Battery
Ok seriously, I need more experiment suggestions.
Step 1: I show you a cheap-ish lab experiment for kids.
Step 2: You give me ideas for what we should try next.
Step 2: You give me ideas for what we should try next.
December 1, 2024 at 2:39 AM
Step 1: I show you a cheap-ish lab experiment for kids.
Step 2: You give me ideas for what we should try next.
Step 2: You give me ideas for what we should try next.