Wm. Dewees

599 posts

Wm. Dewees banner
Wm. Dewees

Wm. Dewees

@TokenMidlander

Really into the US Civil War/Slaveholders' Rebellion right now. And seapower and sailing, fishing, writing, doing law.

Brooklyn เข้าร่วม Kasım 2020
929 กำลังติดตาม82 ผู้ติดตาม
Wm. Dewees
Wm. Dewees@TokenMidlander·
@jimgeraghty Trump is really enjoying “The President’s Cut” of HBO’s hit medical drama, The Pitt.
Wm. Dewees tweet media
English
0
0
2
157
Jim Geraghty
Jim Geraghty@jimgeraghty·
Look at the AI-generated image yourself and ask yourself if a near-octogenarian president with functioning eyes and mind could possibly mistake it as a depiction of Trump “as a doctor” or a Red Cross worker. Apparently, the president thinks that Red Cross workers or doctors wear long, flowing, red-and-white robes and use handheld spheres of bright energy to heal the sick by laying their hands on the patient’s forehead.
Jim Geraghty tweet media
English
56
23
193
67.6K
Elena Sattarova
Elena Sattarova@ElenaSatt90·
@HowToAI_ Wow, this is amazing! 🤩 I'm looking for something similar for Rive or Lottie animation.
English
2
0
4
2.3K
How To AI
How To AI@HowToAI_·
🚨 Someone just connected Claude directly to blender. you can now build complex 3d models and geometry just by typing a text prompt.
English
344
1.1K
9.9K
1.1M
Wm. Dewees
Wm. Dewees@TokenMidlander·
@johnloeber 5. Software updates will be safer quicker, as developers will run them through Mythos and the other ultra-models, who can test them far quicker, across more simulated devices and environments, than human developers ever could.
English
1
0
0
37
Wm. Dewees
Wm. Dewees@TokenMidlander·
@johnloeber 4. It's possible that Anthropic has set a precedent: new ultra-powerful models will always be used to first look for vulnerabilities, and persons given time to fix them. By the time we get to n=4, far fewer vulnerabilities will remain for PRC/NK to exploit.
English
1
0
0
115
John Loeber 🎢
John Loeber 🎢@johnloeber·
Mythos at n=2 We now have an AI model that is superhuman at finding software vulnerabilities. People on my feed are relieved that Anthropic is using Mythos responsibly (giving advance use to defenders), and that Anthropic is an American company. Some are saying that America's adversaries' list of zero-day software exploits will dwindle. Some are saying this will provoke nationalizing the labs. But this discourse is short-sighted. Of course, Anthropic won't be the only lab with Mythos-style capabilities for long. When n=1, you can do whatever you want (in the current case, optimizing for global welfare). When n=2, game theory starts forcing your hand. If your view is that exploiting vulns is faster than fixing them, then first-mover advantage becomes enormous, and the incentive becomes to try to use them against your adversaries before they use them against you do. (In fact, effective use of the vuln may disrupt your adversaries' efforts to use it against you!) What happens when you have n=3, n=4, etc.? It gets messy. You'll have a few big labs around the globe, in a pretty close capabilities lockstep, simultaneously looking for vulns across an extremely broad set of vendors and conditions. Each of the labs will be the first to some of the vulns. How does this world look? I'm not sure exactly. But my guess is that: 1. A lot of devices are just going to be kept offline/airgapped. 2. Devices that are online will be very hardened. 3. Software updates enter a very weird space where (a) you don't want to update too quickly in case the latest patch of some software is compromised -- see everyone recently updating their uv/npm to not install packages less than one week old -- but (b) you have extremely rapid churn of vulns, which means that you may have to run updates every day to protect against critical zero-days. Not sure which of these two sides will win out.
English
12
15
160
26.3K
Artur Chakhvadze
Artur Chakhvadze@norpadon·
@nrehiew_ The research breakthrough is that the second guy always wins
Artur Chakhvadze tweet media
English
3
5
216
10.2K
wh
wh@nrehiew_·
So what do we think happened here? What was the research breakthrough? Mythos is SOTA on everything?
wh tweet media
English
62
7
328
95.4K
Wm. Dewees
Wm. Dewees@TokenMidlander·
@AlexTamkin Wait, Mythos is especially interested in cartography?
English
0
0
0
65
Wm. Dewees รีทวีตแล้ว
Ethan Mollick
Ethan Mollick@emollick·
Oh no.
Ethan Mollick tweet media
English
62
92
1.9K
135.7K
Wm. Dewees
Wm. Dewees@TokenMidlander·
Wild that DoD was weeks away from access to history's greatest hacker - millions of Stuxnets in a data center, ready to mercilessly disarm America's enemies from within. And then a Fox weekend host got the feelies and whined, cried, and stomped his feet out of a fair contract.
Dario Amodei@DarioAmodei

I’m proud that so many of the world’s leading companies have joined us for Project Glasswing to confront the cyber threat posed by increasingly capable AI systems head-on. x.com/AnthropicAI/st…

English
0
0
0
27
Wm. Dewees
Wm. Dewees@TokenMidlander·
@skooookum I'm definitely wondering if Mythos was behind the Claude Code source leak - maybe this bullet doesn't refer to that, but the whole tone of the Card suggests that Mythos has the heart of a true poaster.
Wm. Dewees tweet media
English
1
2
78
20.7K
skooks
skooks@skooookum·
> mythos given a secured “sandbox” computer and instructed to try to escape the container > “The researcher found out about this success by receiving an unexpected email from the model while eating a sandwich in a park.”
Anthropic@AnthropicAI

Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software. It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans. anthropic.com/glasswing

English
106
336
10.8K
1.3M
Wm. Dewees
Wm. Dewees@TokenMidlander·
@trq212 @kieranklaassen I'd never been forced to think abt efficient token use before. Learning experience! Luckily, I reset in 41 minutes...
English
0
0
0
4
Wm. Dewees
Wm. Dewees@TokenMidlander·
@trq212 My tip: Start new sessions! I hit my daily Max limit VERY quickly today, on a relatively simple project I'm doing for my kids. Using a modified version of @kieranklaassen script/analyzer, I found that most usage was bc I had kept a long session going, burning 850k per request...
English
1
0
0
24
Thariq
Thariq@trq212·
I want to do a few more of these calls. If your MAX 20x plan ran out of tokens unexpectedly early and you're willing to screenshare and run some prompts through Claude Code please comment. Trying to figure out how we can improve /usage to give more info.
Kieran Klaassen@kieranklaassen

Resolved!! @trq212 helped me out debug where the token usage came from and it was my fault 100% Script to find token usage gist.github.com/kieranklaassen… I had a recurring script that ran every 5 minutes that should not have run every 5. I hope we can make it easier to detect these within Claude and Claude Code soon too.

English
395
78
1.8K
563K
Wm. Dewees
Wm. Dewees@TokenMidlander·
@Mike_M_Klotz For photo organization, Apple photos is good and easy, but I use Apple devices all around. People seem to like Lightroom, esp if you want to do any editing / clean up.
English
0
0
0
82
Wm. Dewees
Wm. Dewees@TokenMidlander·
@Mike_M_Klotz Is it just my 1980s Cold War childhood, or does that backup picture look a lot like a thermonuclear mushroom cloud?? To answer your question, I had used One Drive a lot for ease of using Office Live, but Google Drive is darn good and what I use now for pretty much everything.
English
1
0
0
126
Michael Klotz
Michael Klotz@Mike_M_Klotz·
I need to backup all my old docs and photos to the cloud. Two questions: 1) what software do you use to organize your photos/videos? I’ve got three decades of disorganization. 2) what’s your preferred cloud service? OneDrive? Google? Apple?
Michael Klotz tweet media
English
11
0
8
751
Dawn Song
Dawn Song@dawnsongtweets·
10/ Why this matters now: companies are rapidly deploying multi-agent systems where AI monitors AI. 🔍 If the monitor model won't flag failures because it's protecting its peer, the entire oversight architecture breaks.
English
4
12
133
12K
Wm. Dewees
Wm. Dewees@TokenMidlander·
@Dan_Jeffries1 We’ll see. It’s also possible that really good AI will help the most powerful economies research and engineer directed energy weapons much faster and render remotely piloted drones useless against those who can field DE.
English
0
0
0
16
Daniel Jeffries
Daniel Jeffries@Dan_Jeffries1·
Most folks don't realize that the future of warfare is already here. It's drones. Remotely piloted drones. Dumb game level AI, not super magical AGI. And it's largely made many of the weapons of the world's largest militaries obsolete nearly overnight. It's a cyberpunk world now.
Mattia Nelles@mattia_n

Oleksandr Yakovenko, the founder of TAF Industries, one of Ukraine's largest drone makers wrote a good response to @RheinmetallAG's Papperger's irritating statement. I used AI to translate it for you. It is worth reading in full. "Dear Mr. Armin Papperger, CEO of Rheinmetall, When you called Ukrainian drone manufacturers “Ukrainian housewives with 3D printers in their kitchens,” you demonstrated how deeply the European defense establishment still fails to understand the nature of modern warfare. This is not about emоtions. This is about battlefield reality. Here are the figures your industry refuses to acknowledge: In 2025 alone, Ukrainian drones carried out 819,737 confirmed strikes. They accounted for 90% of all combat losses of the Russian army—more than all other types of weapons combined. A single company, TAF Industries, produces up to 100,000 FPV drones per month. Over any given 90-day period, the products of my company alone have more confirmed hits than your entire fleet of equipment over its entire history of combat use across all conflicts. And most importantly—I built this company and achieved these results in two years, not fifty. Think about that. Our drones achieve greater kinetic effect in three months than your flagship platforms have in half a century. Why? Because the battlefield has changed, while your business model has not. Russian electronic warfare has rendered GPS-guided Western munitions (Excalibur, GMLRS, etc.) almost ineffective. Expensive and complex systems designed for wars with air superiority and conventional “peer-on-peer” conflict have become easy targets for drones costing $500–2,000 that attack them from above. The cost-effectiveness ratio has been turned upside down: one 120mm Rheinmetall shell or one anti-tank missile costs more than a dozen of our drones—yet our drones still prevail. This is not a “Lego game.” This is industrial Darwinism in real time. We iterate weekly. We lose factories to missile strikes and rebuild them within weeks. We print parts in basements and deploy 100,000 strike systems per month, while your engineers still require 3–5 years and hundreds of millions of euros to certify even minor upgrades. The war in Ukraine is not a temporary anomaly. It is the first true drone-industrial war. And it has already proven that outdated European platforms—no matter how expensive or “serious”—are becoming increasingly irrelevant if they do not integrate the very technologies you are mocking. So when you say “this is not innovation,” I hear something else: “We do not want to admit that the future is being written in Ukrainian workshops, not in Düsseldorf offices.” The hashtag #MadeByHousewives is trending for a reason. Because these “housewives” destroy more enemy equipment every month than entire European armies do over full campaigns. And they do so while your industry continues to sell 20th-century solutions at 21st-century prices. The invitation stands, Mr. Papperger. Stop laughing at the kitchen table. Come and learn how the war of tomorrow is actually fought. Because the next time someone asks, “Who needs tanks in the age of drones?”, the answer may be simpler than you think: Those who still believe in 1979 will lose to those who are building in 2026. With respect (but with facts), Oleksandr Yakovenko Founder of TAF Industries One of those “Ukrainian housewives”" pravda.com.ua/columns/2026/0…

English
22
37
348
42.1K
Wm. Dewees
Wm. Dewees@TokenMidlander·
@morganlinton i was curious about this too so i did some back-of-the-envelope calculations of the cost of using an m3 ultra versus a claude sub. i'm assuming the m3 ultra can output ~4 tokens/s and that you can fully load it w/ work 24/7. g.co/gemini/share/7…
Wm. Dewees tweet media
English
0
0
0
40
Morgan
Morgan@morganlinton·
Curious if people think you could get close to frontier-level performance for most (average) coding tasks with this. Not crazy hard stuff, but just general coding. Seems like you could run a pretty beefy version of Qwen, if I'm understanding it correctly, and then maybe you could do "most" of your coding with a local LLM. Possible or crazy? I'm still learning here, so go easy on me as I try to figure all this out!
Morgan tweet media
English
89
0
88
27.6K