GoFly

@Drew_OnAI yes，you are right. GPT 5.5 finish abaqus cae simulation

@GoGoFly23 gpt 5.5 meshing a bullet hole? cool story for a demo, terrible for real engineering i'd bet my paycheck it hallucinates the stress points

Codex联合Abaqus完成由AI驱动CAE仿真，从悬臂梁的静力学分析、圆孔板的模态分析、再到子弹穿过钢板的动力学分析，都可以由Codex的GPT 5.5来完成。 从三维模型的建立、装配、材料属性的定义、载荷定义、边界条件、网格划分、Job的提交、后处理这仿真的一切，都可以交给AI来完成。 #GPT5.5 #GPT #AI #CAE #CAD #Abaqus #Agent

@penguinapplied 不，现在 AI 还不能取代人类

解析人材はもうAIに取って変わるだろうなあ 如何にシミュレーションでやられていないところを探すか

@t_kun_kamakiri Awsome

すごそう

@shimura_werdex 非常に驚くべきことであるが、codexは完全に自動的にabaqusを駆動しているのである。

これどうゆう体験？操作体験にCodexがどう関わってんの？

@with_ai_hisyo AI時代においては、確かにあらゆる業務をAIに任せることが可能である。

直感的にシミュレーションが自然言語でできるようになったら、ものづくりのハードルがますます下がる。 素晴らしい時代だ

@CC98842318 确实非公开的确实没办法

@GoGoFly23 这里面材料参数大部分非公开的，比较难获取

@Jeff_John_sta 我觉得是的

@GoGoFly23 哈哈！这些都只是课堂上学CAE的入门例子，除了课堂上用，连考试和面试都不用，除非送分。只能说AI非常外行，真正的，比如仿真所有的试验数据库AI能不能不交订阅费就帮咱搞到手？AI能不能把各零件的微小间隙给帮咱找到？只能说AI的泡沫的确灰常大。

@artemihq Of course, we still need to be worried, AI will make many mistakes at present.

@GoGoFly23 i'm not sure engineers have nothing to worry about

@osaAtwi AI is all

Most "3D datasets" are useless for training engineering AI Pretty meshes. No construction history. No simulation. No manufacturing context. No assembly constraints. A model trained on that produces geometry that is useless for real world applications. Not all 3D models are created equal. What's actually needed: - Feature trees: the construction sequence that produced the part, not just the final geometry - Simulation-validated geometry: load cases, stresses, what the part can actually withstand - DfM metadata: how the part is meant to be manufactured, and the constraints that implies - Geometric and assembly constraints: how the part connects to other parts in assemblies Anything less is nice-looking meshes with no engineering content. If clean parametric 3D data is your bottleneck, let's talk.

Dynamic simulation of abaqus bullet penetrating steel plate, static simulation analysis of cantilever beam and tensile plate with circular hole, modal analysis are carried out by using GPT 5.5 model of codex. AI is used to perform tedious simulation setup workflow, only communicating with codex text throughout, and the rest is handed over to codex to complete the simulation work. text to cae. Will AI replace simulation engineers? Or will simulation engineers integrate AI into their own workflow? #gpt5.5 #codex #cae #agent #model #openai

@7GHz @MattiaFiumara You're right, it's not accurate.

@MattiaFiumara @GoGoFly23 実際に生成されたものですが正しくありません。 ボリュームが片方にしか効かずに壊れてしまいます。

GPT Image 2 测了一组电路图生成效果，包含原理图、PCB、PLC、和微电子版图。 整体工程感、布局逻辑和图纸完整度让我这个外行看的是一愣一楞的 电气工程师会失业吗？ 各位电气大佬来看看 GPT Image 2 生成的电气原理图如何？ #GPTimage2 #ai #GPT

@haomingzi445 @MEnough4 有可能是的

@GoGoFly23 @MEnough4 我觉得是因为ai很容易出错而且这方面的测试成本比较高

@Kodolskiltava So how can AI better help electrical engineers?

@GoGoFly23 If me, as an engineer, would open an article with these illustrations, it's quite possible that I'd close it without reading. At first they look lehot and then ... then it's clear that it's just AI shit. Thechical drawings are techical for a reason. They should make sense.

@he22ho22 @seasonedslacker maybe

@GoGoFly23 @seasonedslacker It's LLM. It should not be used for engineering, it will never be good because it's language model ffs

@losslandscape Yes, we need to give AI more precise prompts to constrain it to ensure more accurate output, but AI will indeed

One of my personal tests for new models is to ask it to generate simple ham radio circuits. Image 2 is the best yet, but it still gets a lot wrong. Eventually, though.

@AssistedEvolve It is true that using code to correspond to circuit diagrams may be more accurate, but how should it be done in practice?

@GoGoFly23 No this will not disrupt anything, its the wrong approach. However circuits as code is totally tractable and progress is moving fast in that direction

@tobjectpascal Your worries are very necessary. At present, the consideration of AI is not very comprehensive. Most of the circuit design also has many errors, but we can optimize their design through prompts.

@GoGoFly23 Yes and no If anything it should help engineers, drive down cost and save the environment as less versions of the board will ge needed to fix issues My question does it take heat into consideration if a chip is outputting lots of heat, we don't want capacitors directly nexttoit

@tobjectpascal @hackradios now？

@hackradios @GoGoFly23 For now....