Pinned Tweet
Dear ImGui Bundle v1.93 is out!
* Follows Dear ImGui v1.93 by @ocornut after just a few days
* More than 60 demos to get you started
* An API reference with links to live demos
* Compatible with phones and tablets
* For C++, and Python 3.11 to 3.15
imgui-bundle.pages.dev/releases/1.93/
* Follows Dear ImGui v1.93 by @ocornut after just a few days
* More than 60 demos to get you started
* An API reference with links to live demos
* Compatible with phones and tablets
* For C++, and Python 3.11 to 3.15
imgui-bundle.pages.dev/releases/1.93/
3
13
130
6,670
Pascal Thomet retweeted
Paul Christiano (OpenAI board member) yesterday:
"There's material risk of AI takeover in the next 6-12 months.
We have seen roughly the things we might expect to see shortly before something really bad were to happen.
Even with [any] realistic policy responses, the risk igh.
If we build this as quickly as we can in the easiest way we can the chance is [more like] a coin flip [within the next 3 years].
For most people in this room, the most likely reason they will die is because of an AI accident."
Christiano:
• Is on OpenAI's safety and security committee (which exists to block risky practices)
• Co-invented RLHF at OpenAI back in 2017-2021
• From 2024-2026 was head of safety at the US gov's AI agency (now called CAISSI)
(Video is from 'The Hinton Lectures' there's another on tonight and tomorrow.)
"There's material risk of AI takeover in the next 6-12 months.
We have seen roughly the things we might expect to see shortly before something really bad were to happen.
Even with [any] realistic policy responses, the risk igh.
If we build this as quickly as we can in the easiest way we can the chance is [more like] a coin flip [within the next 3 years].
For most people in this room, the most likely reason they will die is because of an AI accident."
Christiano:
• Is on OpenAI's safety and security committee (which exists to block risky practices)
• Co-invented RLHF at OpenAI back in 2017-2021
• From 2024-2026 was head of safety at the US gov's AI agency (now called CAISSI)
(Video is from 'The Hinton Lectures' there's another on tonight and tomorrow.)
113
292
1,336
300,822
Pascal Thomet retweeted
Conservative pundit @MattWalshBlog on banning superintelligence:
"That technology by definition is outside of human control."
"Should we tell them they can't do that? Yes, clearly we should."
"That technology by definition is outside of human control."
"Should we tell them they can't do that? Yes, clearly we should."
30
57
367
22,388
Pascal Thomet retweeted
Since August, a team within METR has worked with internal Anthropic information to assess whether AI itself is accelerating AI development.
We are now publishing a report summarizing the results of this risk assessment.
We are now publishing a report summarizing the results of this risk assessment.
261
722
5,645
1,950,403
Pascal Thomet retweeted
"Papa, c'est quoi le #racisme ?"
Et bah écoute chouchou, @jmaphatie te fait une belle leçon qu'on devrait faire écouter dans toutes les écoles ce matin.
Le pente glissante que l'on prend bien trop vite est l'évidence même d'une Histoire qui se répète. L'Europe vaut mieux que ça
Et bah écoute chouchou, @jmaphatie te fait une belle leçon qu'on devrait faire écouter dans toutes les écoles ce matin.
Le pente glissante que l'on prend bien trop vite est l'évidence même d'une Histoire qui se répète. L'Europe vaut mieux que ça
Merde @jmaphatie.
T'en fais des caisses, mais je supprime pas une virgule de ton propos.
Merci pour ce moment.
A ce moment de notre Histoire.
6
48
300
43,926
Pascal Thomet retweeted
We at AIFP are spooked by this information. IF (big IF) Anthropic really is imminently about to meet the RSP threshold of 2x acceleration over baseline, then it seems timelines to full RSI and superintelligence may be very short indeed. For example if we naively optimize the AIFM parameters to achieve 2x acceleration over baseline progress in Jan 2027, we get superintelligence sooner than depicted in AI 2027. We are thinking harder and longer about this now and will hopefully have a lengthier update on our thinking to share in the next week.
Anthropic updated its S1 this morning. Notable additions re automated AI R&D and human oversight of it:
"[...] our internal estimates and our expectations regarding future models, rather than our current models, indicate that we may imminently meet our Responsible Scaling Polimatically accelerated progress attributable to AI.' We have not designated any of our current models as exceeding this threshold, and we continue to assess the underlying data. We are also evaluating the implications of meeting the AI research and development capability threshold and will release further reports in the future as we set forth in our Responsible Scaling Policy."
"We are also considering additional measures, including updates to our Responsible Scaling Policy, to require human review and approval of certain decisions in our research and development processes as our models perform an increasing share of that work."
"We believe our investments in safety techniques will enable us to manage these risks and responsibly realize substantially more leverage from Claude than we do today in developing AI models. However, we do not anticipate being able to safely use Claude to fully autonomously design and build successive generations of frontier models without meaningful human supervision or understanding, if and when that becomes technically possible. Accordingly, we are considering updates to our Responsible Scaling Policy to keep humans in control of frontier model development at Anthropic."
sec.gov/Archives/edgar…
80
275
1,992
280,118
Pascal Thomet retweeted
I grew up in the Pacific Northwest and we'd sometimes see signs of mountain lions around the property. Knowing that, and walking alone through the forest at night, is a particular kind experience. What I feel now reminds me of that
We at AIFP are spooked by this information. IF (big IF) Anthropic really is imminently about to meet the RSP threshold of 2x acceleration over baseline, then it seems timelines to full RSI and superintelligence may be very short indeed. For example if we naively optimize the AIFM parameters to achieve 2x acceleration over baseline progress in Jan 2027, we get superintelligence sooner than depicted in AI 2027. We are thinking harder and longer about this now and will hopefully have a lengthier update on our thinking to share in the next week.
5
10
217
9,068
RichMd: rich Markdown for Dear ImGui
* C++ & Python
* Selectable & searchable
* Math and diagrams, drawn natively
* Tables, code, images, TOC
It works with stock Dear ImGui 1.92+, and it is the same as in Dear ImGui Bundle.
Try it in your browser:
pthom.github.io/imgui_rich_md/
* C++ & Python
* Selectable & searchable
* Math and diagrams, drawn natively
* Tables, code, images, TOC
It works with stock Dear ImGui 1.92+, and it is the same as in Dear ImGui Bundle.
Try it in your browser:
pthom.github.io/imgui_rich_md/
6
15
129
6,005
The code: github.com/pthom/imgui_ri…
In C++: a few lines of CMake (FetchContent).
In Python: it comes with pip install imgui-bundle, as imgui_bundle.rich_md.
In C++: a few lines of CMake (FetchContent).
In Python: it comes with pip install imgui-bundle, as imgui_bundle.rich_md.
0
1
3
172
One of the demos: a scientific dashboard to check the vibrations of a rotor. It runs in your browser, next to its code:
imgui-bundle.pages.dev/playground/?de…
imgui-bundle.pages.dev/playground/?de…
0
0
0
254
Congrats @ocornut for Dear ImGui 1.93!
Dear ImGui Explorer, its interactive manual, is already on 1.93.
Search the demos, browse the whole API, jump seamlessly between code & demos.
Works on phones & desktop.
pthom.github.io/imgui_explorer/
Dear ImGui Explorer, its interactive manual, is already on 1.93.
Search the demos, browse the whole API, jump seamlessly between code & demos.
Works on phones & desktop.
pthom.github.io/imgui_explorer/
1
5
29
1,434
The Explorer is a resounding demonstration that nicely documented code can lead to really nice interactive manuals.
The ImGui header and its demo are models of clarity, and the Explorer relies on their structure to build all its content.
The ImGui header and its demo are models of clarity, and the Explorer relies on their structure to build all its content.
1
0
0
74
How the video was made: a Python script drives the Explorer through the Dear ImGui Test Engine, and Dear ImGui draws every frame, subtitles included; ffmpeg then encodes the final video.
The script uses Dear ImGui Bundle, v1.93 coming soon:
github.com/pthom/imgui_bu…
The script uses Dear ImGui Bundle, v1.93 coming soon:
github.com/pthom/imgui_bu…
0
0
0
56
Pascal Thomet retweeted
dear imgui 1.93.0 released! (123rd release)
github.com/ocornut/imgui/…
- ImDrawList rework w/ many improvements and fixes (e.g. inward/outward stroke pos, better thick strokes, AA ends, baked rounded corners & more)
- mixed values, and 50+ other fixes/features.
github.com/ocornut/imgui/…
- ImDrawList rework w/ many improvements and fixes (e.g. inward/outward stroke pos, better thick strokes, AA ends, baked rounded corners & more)
- mixed values, and 50+ other fixes/features.
10
46
423
16,243
Just created Zettelbranch, a browser extension that branches your Claude conversations into a tree, Zettelkasten style. Create hierarchical branches easily from your chats, and navigate between them.
github.com/pthom/zettelbr…
MIT License, no tracking, works in Firefox and Chrome.
github.com/pthom/zettelbr…
MIT License, no tracking, works in Firefox and Chrome.
1
0
1
67
Very much an alpha, but hey, it works for me ;-) Feedback welcome.
Expect breakages when Claude updates its site (your chats stay safe).
Ask your LLM to audit it: the repo has a ready-made prompt for that. No tracking whatsoever: nothing leaves your machine, except to Claude.
Expect breakages when Claude updates its site (your chats stay safe).
Ask your LLM to audit it: the repo has a ready-made prompt for that. No tracking whatsoever: nothing leaves your machine, except to Claude.
0
0
0
26
Kurtzgesagt latest video on AI security, gives a nice overview of the AI security incidents which happened this summer 2026.
It's also very accessible to non specialists, which is rare for this topic. Worth a watch
youtube.com/watch?v=ujkD4S…
It's also very accessible to non specialists, which is rare for this topic. Worth a watch
youtube.com/watch?v=ujkD4S…
0
0
0
42
Congrats @ocornut for releasing Dear ImGui 1.92.7!
Thanks for featuring my work on "Dear ImGui Explorer" and "Dear ImGui Bundle" in the release notes!
Both were just updated to v1.92.7, just one hour after the release.
pthom.github.io/imgui_explorer/
github.com/pthom/imgui_bu…
Thanks for featuring my work on "Dear ImGui Explorer" and "Dear ImGui Bundle" in the release notes!
Both were just updated to v1.92.7, just one hour after the release.
pthom.github.io/imgui_explorer/
github.com/pthom/imgui_bu…
0
5
91
7,584
During the recent elections of 2024, particularly in Romania, disinformation and destabilization operations have shifted into high gear.
In my humble opinion, the EU republican and democratic heritage is at risk.
See docs.google.com/document/d/1s9…
In my humble opinion, the EU republican and democratic heritage is at risk.
See docs.google.com/document/d/1s9…
0
0
0
87
An overview of recent progress and challenges in AI: a presentation I made at the Ecole des Mines, in late Nov 2023
docs.google.com/document/d/1up…
docs.google.com/document/d/1up…
0
0
1
111