Draxler

170 posts

Draxler

Draxler

@CausalEngineer

Machine Learning and AI Engineer

Katılım Aralık 2025
329 Takip Edilen12 Takipçiler
Draxler
Draxler@CausalEngineer·
Using fear of AI as a marketing strategy is a kind of Frankenstein anxiety. It may give your company a market advantage, but it also narrows the real possibilities of what people could build with your models.
Dustin@r0ck3t23

Jensen Huang just told every AI leader in the room to grow up. Stop scaring the public with science fiction. Start communicating like the weight of civilization is on your shoulders. Because it is. Huang: “AI is not a biological being. It is not alien. It is not conscious. It is computer software.” That single statement dismantles half the panic surrounding this industry. The mainstream conversation is dominated by people projecting human malice onto math. Alien consciousness onto code. Existential dread onto a software architecture we built, we trained, and we can read. Huang: “We say things like, ‘We don’t understand it at all.’ It is not true. We understand a lot of things about this technology.” When builders tell the public they don’t understand their own creation, the public hears threat. The state responds with control. That is already happening. Palihapitiya asked Huang what he would have told Anthropic during their regulatory clash with the Department of Defense. Huang didn’t attack the technology. He attacked the communication. Huang: “The desire to warn people about the capability of the technology is really terrific. We just have to make sure that we understand that the world has a spectrum, and that warning is good, scaring is less good because this technology is too important to us.” Warning shows risks, mitigation, why upside overwhelms downside. Scaring says we might be building something that destroys us and we can’t stop it. One builds trust. The other invites regulation written in panic. Huang: “To say things that are quite extreme, quite catastrophic, that there’s no evidence of it happening, could be more damaging than people think.” Projecting catastrophe without evidence is not caution. It is sabotage. When your technology is embedded in national defense, the financial system, and healthcare infrastructure, your words carry structural weight. If the architects act terrified of their own product, the response is predictable. Governments step in. They restrict. They seize control of something they don’t understand because the builders told them to be afraid. Huang: “There was a time when nobody listened to us, but now because technology is so important in the social fabric, such an important industry, so important to national security, our words do matter.” Most tech founders have not internalized this. You are no longer a startup founder disrupting an industry. You are running infrastructure that nations depend on. Your statements move policy. Your framing shapes legislation. Your tone determines whether governments treat you as partner or threat. Huang: “We have to be much more circumspect, we have to be more moderate, we have to be more balanced, we have to be far more thoughtful.” Huang did not ask for silence. He asked for precision. The leaders who cannot tell the difference will not be leading for long.

English
0
0
0
16
Draxler
Draxler@CausalEngineer·
@_arohan_ People might ride horses again if oil ever hits $200. And if manual coding becomes rare, those who truly understand code review will be like elite horse riders in the age of cars.
English
0
0
1
28
rohan anil
rohan anil@_arohan_·
Writing code character by character has become equivalent to riding horses. You can instead learn to drive a car to get places. And occasionally do what people in middle ages did when claude is down.
English
4
1
14
1.1K
Draxler
Draxler@CausalEngineer·
@mandylu I remember when everything was handwritten.
English
0
0
0
34
Mandy Lu
Mandy Lu@mandylu·
remember when everything was human-generated?
English
29
1
56
2.1K
Draxler
Draxler@CausalEngineer·
@sundeep The only limits left are imagination and compute, and both keep stretching as chips improve and agent workflows mature.
English
0
0
1
38
sunny madra
sunny madra@sundeep·
I have no problem staying up until the middle of the night every night just building these days... there is literally no limit on what can be done...
English
10
3
56
1.9K
Draxler
Draxler@CausalEngineer·
@Yuchenj_UW “Claude, you are a seasoned tax auditor. File my 2025 taxes and do not hallucinate a single deduction.”
English
0
0
0
22
Draxler
Draxler@CausalEngineer·
@Yuchenj_UW Codex seems to prefer being a ghost software engineer, like a ghostwriter behind a great book. A bit like Bourbaki, but for software.
English
0
0
0
104
Yuchen Jin
Yuchen Jin@Yuchenj_UW·
I noticed something interesting: Claude Code auto-adds itself as a co-author on every git commit. Codex doesn’t. That’s why you see Claude everywhere on GitHub, but not Codex. I wonder why OpenAI is not doing that. Feels like an obvious branding strategy OpenAI is skipping.
English
193
30
1.5K
133.5K
Draxler
Draxler@CausalEngineer·
Software engineering is basically an open-book exam. When you face a new problem or a new programming language, you rely on docs, references, tools, tests, and the harness. That is much closer to real software work than a closed-book setting.
Lossfunk@lossfunk

🚨 Shocking: Frontier LLMs score 85-95% on standard coding benchmarks. We gave them equivalent problems in languages they couldn't have memorized. They collapsed to 0-11%. Presenting EsoLang-Bench. Accepted to the Logical Reasoning and ICBINB workshops at ICLR 2026 🧵

English
0
0
0
35
Draxler
Draxler@CausalEngineer·
@trq212 That’s why I don’t spend compute telling Claude “thanks”. I’d rather save that for Anthropic’s engineers, at least until compute feels as cheap and invisible as bandwidth on fiber.
English
0
0
3
392
Thariq
Thariq@trq212·
an increasingly large part of the job of an engineer is deciding how much compute to spend on a problem
English
138
66
1.6K
69.1K
Draxler
Draxler@CausalEngineer·
@himanshustwts It’s partly a data curation and style modeling problem. We can train TTS models on high-quality, context-appropriate speech data to better capture natural prosody, while treating highly performative sources like Twitch speech as a biased or noisy distribution.
English
1
0
1
35
himanshu
himanshu@himanshustwts·
there is lot of scope to improve current text-to-speech models like dude optimizing just over "expressiveness" will not lead towards "naturalness". and there is a tradeoff. being “natural” is not a scalar property of a waveform but more likely a conditional distribution over realizations given persona and situation.
English
1
1
21
826
Draxler
Draxler@CausalEngineer·
@slow_developer Some people get attached to AI because it appears to offer something humans often don’t, like constant availability, patient responses, and little visible judgment. For people who feel uneasy with human unpredictability, that can become emotionally compelling.
English
0
0
2
198
Haider.
Haider.@slow_developer·
i still don't understand the attachment people have to LLMs it is a computer, not a friend. for those who missed the older models in this way, it seems many are unhappy with openai's current direction i need a research assistant, so i don't care much about that
English
117
7
105
12K
Draxler
Draxler@CausalEngineer·
@andrewchen That’s what happened with computers in the 90s. They made people more productive, but they also increased the amount of work people took on. Back then we were building the internet. Now we’re building a new world of agents.
English
0
0
0
262
andrew chen
andrew chen@andrewchen·
AI is supposed to save me time, but now I find myself building stuff all evening and weekend and it's actually increasing my time in front of the computer WTF
English
446
124
2.1K
180.8K
Draxler
Draxler@CausalEngineer·
@fhinkel Gemini already works well because it is tightly integrated with Google’s own stack. The real gap is with third-party products. It feels strongest inside Google Calendar, Meet, Gmail, and Drive, and much weaker once you leave Google’s ecosystem.
Draxler tweet media
English
0
0
0
35
Franziska Hinkelmann, PhD
AI is going to take our jobs! The AI: I've successfully rescheduled your 1-1 with Bob to start at 2:35 PM. The rescheduled meeting: no title, no guest. Guess I'll just meet with myself. Or I should probably file a bug. "AI, file a bug for me..."
Franziska Hinkelmann, PhD tweet media
English
10
1
25
1.6K
Draxler
Draxler@CausalEngineer·
@Yuchenj_UW I disagree… Frontier labs can get bigger, but they cannot absorb every new idea or solve every emerging bottleneck, from inference to cloud computing. As long as humans keep generating both, startups will still matter.
English
0
0
0
1.5K
Yuchen Jin
Yuchen Jin@Yuchenj_UW·
Some people at frontier AI labs told me they believe startups are over. OpenAI, Anthropic, Google, xAI will absorb every industry as AGI nears. Coding today, science, medicine, and finance next. Then everything else. If they’re right, that’s a pretty boring end of the world.
English
537
162
3K
880.5K
Draxler
Draxler@CausalEngineer·
@samhogan Precision can still be enforced through code review, testing, and verification, whether the code comes from a human or an AI. The real issue is that AI can generate code at a scale that becomes difficult to review, especially in larger projects.
English
0
0
0
70
Sam Hogan 🇺🇸
Sam Hogan 🇺🇸@samhogan·
Building good software requires precision. Anyone who has worked on systems at scale will agree with this. You cannot use an LLM to generate code with precision. draw you’re own conclusions
English
6
0
17
1.4K
Draxler
Draxler@CausalEngineer·
Gamers are fiercely defending artistic identity against something like optional DLSS 5, but the history of graphics is full of the same reaction to almost every visual leap since the move to 3D. First they resist it, then it becomes the new normal.
NVIDIA GeForce@NVIDIAGeForce

Announcing NVIDIA DLSS 5, an AI-powered breakthrough in visual fidelity for games, coming this fall. DLSS 5 infuses pixels with photorealistic lighting and materials, bridging the gap between rendering and reality. Learn More → nvidia.com/en-us/geforce/…

English
0
0
0
36
Draxler
Draxler@CausalEngineer·
Podcasts sit in an awkward middle ground. Some people want the visual side for reactions, clips, and on-camera moments, while others just want the audio while driving or falling asleep. I can see the product logic behind splitting those use cases, even if the UX still feels fragmented sometimes. I use both apps myself.
English
0
0
0
32