Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 (@btnelson) header
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 (@btnelson) profile picture

Technology and stuff

Amsterdam, The Netherlands
Joined December 2008
Photo by Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 (@btnelson)Photo by Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 (@btnelson)Photo by Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 (@btnelson)Photo by Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 (@btnelson)Photo by Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 (@btnelson)Photo by Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 (@btnelson)Photo by Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 (@btnelson)Photo by Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 (@btnelson)Photo by Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 (@btnelson)Photo by Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 (@btnelson)
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
show this video to anybody that is an old school programmer that hates AI this is one of the best videos I’ve seen all month.
1,138
6,581
1,361
47,677
4,405,941
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
Moore's law: available compute increases exponentially Dario's law: required compute increases exponentially
7
4
82
6,313
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
5,000 years of world history on a single timeline. Zoom in. You could spend 30 minutes exploring this.
424
4,679
241
19,616
3,296,006
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
A programmer and a software engineer aren’t the same. I’ve known many people who made a living writing code, but they were always a cog in a larger process they never understood. Frontier models can do that job today. 1000x better. Software engineers can write code, but they always understood code was a means to an end. Software engineering is alive and well.
99
34
11
452
32,287
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
To offend a strong man, tell him a lie, To offend a weak man, tell him the truth — Marcus Aurelius
Readers added context they thought people might want to know
This quote is falsely attributed to Marcus Aurelius and does not appear in the Meditations or his verified writings. classics.mit.edu/Antoninus/medi… stoicshelf.com/posts/fake-mar… stoichandbook.co/marcus-aureliu…
61
1,343
78
11,689
207,510
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
me: how do i open pdf claude: You're right to ask. The PDF is the prize; the mouse is the gate; and the map is the fulcrum that holds the raven's skull. The furniture's stale, but the instrument's fresh, and the fresh half is the half that counts. I deleted your home directory.
229
2,012
124
30,412
771,736
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
The executives said the boardroom Wi-Fi “feels slow.” Speed test: 487 Mbps. Apparently 487 Mbps has bad energy. So I created a new network: EXECUTIVE_PRIORITY_5G Same access points. Same internet connection. Same VLAN. I just gave it a more expensive name. Connected the CEO. He opened a website and immediately said: “Oh wow, this is way faster.” By lunch, five VPs had asked for access. I told them executive bandwidth is limited and I can only approve users with a genuine business requirement. They are now competing for access to Wi-Fi they already had. Performance is unchanged. Satisfaction is through the roof.
346
1,712
158
42,546
1,165,061
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
OpenAI's Noam Brown says air-gapping the computers may not stop a misaligned AI, because two air-gapped machines can still talk by running a CPU hot and reading the temperature change "But I think the major takeaway from the incident is that people underestimated the AI. And we never want to be in a situation again where we underestimate the AI. It's a weird world, because AI progress is so fast that people are consistently underestimating the AI." "So to be in a situation where you don't underestimate it again, when it comes to safety and alignment, you have to have a very, very, very high bar." "You could even go as far as to say, "Well, we should air gap the computers." And I'm not convinced that that would be sufficient." "There are studies, and this is mostly academic, where you can have two computers next to each other that are air-gapped and they're still able to communicate with each other because they have temperature sensors." "One of them is able to run their CPU really hot, and then the other one can actually detect the temperature change, and then that actually gives them a mechanism to communicate." _________ Link and more key quotes from OpenAI's safety related conversations: firesidealpha.substack.com/p…
Noam Brown reveals OpenAI is already watching chain-of-thought monitorability degrade as models get better at controlling what they show "And this is one major concern, and we're already seeing signs that chain of thought monitorability is degrading, for various reasons." "We're trying to figure out exactly why, because we want to reverse the trend. But we're seeing that the model is becoming better able at controlling its chain of thought." "So this is a problem, because you could have a situation where the model understands what chain of thought is and that people are observing it." "And eventually they will, because this is all in the pre-training data. The idea of chain of thought monitoring has been around long enough that it's in the pre-training data, they're aware of it, but they're not actually able to control their chains of thought." "If we reach a point where they're actually able to recognize, "Oh, I am being observed, I want to think these bad thoughts in a way that is not observable to my monitors," and then they're able to actually do that, then there's a problem." _________ More key quotes from OpenAI's safety related conversations: firesidealpha.substack.com/p…
864
199
1,000
2,284
4,946,622
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
MIT published a brutally honest report on what AI is doing to students. A committee of professors and students spent five months studying how AI changed learning on campus, and the findings read like a warning to every university on the planet. Study groups are disappearing. Office hours are emptying out. Problem sets and take-home exams no longer prove anything, because AI can produce credible solutions to almost any written assignment in the undergraduate curriculum. Students who lean on chatbots lose mastery and confidence, and some slip into what the report calls cognitive surrender, reaching for AI at the first hint of struggle. The numbers are rough. 46 percent of surveyed MIT undergrads use LLMs daily. 90 percent worry about their own overreliance. Undergrads who feel AI makes them replaceable now outnumber those who feel it makes them capable. The committee's answer surprised me. They refused to fight AI with surveillance. The report calls AI detectors unreliable, says lockdown browsers feel like spying, and warns that policing students builds a classroom atmosphere of mutual distrust. Instead, MIT wants to rebuild education around the things AI can't replace. That means oral exams, semester portfolios, in-person project work, and a required social component in every subject. The report even floats the idea of rethinking grades entirely, since without a GPA to optimize, much of the incentive to cheat with AI evaporates. The committee warns professors against replacing undergrad research assistants with AI agents just because they're cheaper, because a university exists to grow people, not output. The most famous tech school on earth admitted the machines broke its way of teaching. Its answer is more humans, not more software.
481
8,105
1,089
24,778
3,467,022
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
Saying that a false thing is true does not make that thing true, no matter how many times or how many people say it.
20
33
260
9,332
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
AI rewards expertise (at least for now). Expertise lets you judge AI output quality and find the shape of the jagged frontier quickly. It also gives you more options for how to try to improve quality by knowing what changes to ask for. Non-experts are often stuck with defaults.
82
80
5
706
34,885
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
Visual Proof: 1/4 + 1/16 + 1/64 + … = 1/3 three colors, three equal shares.
19
61
2
762
54,588
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
Now just 1.5 hours into this mini project I have replaced ALL my Chrome extensions with my own vibe coded ones for safety After another one of my extensions after updating suddenly wanted to read my entire browser history (probably to sell off to an ad company) My own Chrome extension is called 🚀 SuperLevels and now has: 🗑️ Tab Cleaner, auto closes tabs after inactivity (with host-based exclusions) 🍪 Cookie Editor, lets you nuke all cookies or edit any 🔀 Redirect Tracer, lets you see any redirects a site took you through 🌗 Dark Mode, changes websites to dark, per site or all sites 🎚️ X Dim, changes X background back to dark blue 🎶 Music Finder, records music in a tab and submits to find the name of the song 🗺️ Add back a Maps link to Google in EU (they hide it!) 🌠 Add back a View Image link to Google Images in EU (they hide it too!) I've deleted all my other extensions (except uBlock Origin for ad blocking!), much much safer because now I control the source code!
Okay it's been 18 minutes and I've now vibe coded and replaced 3 more extensions and made it all into one super extension called 🚀 SuperLevels It has a 🗑️ Tab Cleaner, auto closes tabs after inactivity (with host-based exclusions) 🍪 Cookie Editor, lets you nuke all cookies or edit any 🔀 Redirect Tracer, lets you see any redirects a site took you through I can read and check the source code, so it's much safer than these extensions that keep being bought by blackhat companies to add malware and steal your data or worse hack you
29
54
36
1,326
215,739
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
Remember the paper saying ChatGPT Health undertriaged 52% of medical queries? Turns out the study methodology was massively flawed. In real world conversations, GPT-5.4 correctly recommends immediate care in emergency cases more than 99% of the time. Healthcare AI isn't perfect and there are open questions about how to evaluate LLM-based tools. But always worth remembering that there are groups that have an incentive to produce a sensational, AI-skeptical headline by cloaking suspect science in academic authority.
🆕 @NatureMedicine How does ChatGPT Health do for appropriately triaging a person as to whether to go to the emergency room or stay home? nature.com/articles/s41591-0… Not very well. Under-triaged 52% of case vignettes that are considered gold-standard emergencies, like diabetic ketoacidosis or impending respiratory failure
Readers added context they thought people might want to know
This study relies on multiple choice questions and does not allow followup questions, which does not reflect real world performance well. When followup questions are allowed from the model, performance increases drastically. x.com/thekaransingha…
25
51
10
605
70,626
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
It's odd that people haven't realized that prompting techniques are actually meta-thinking techniques.
18
22
3
198
8,254
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
A truly great human is still better than any AI agent at most complex, judgment-intensive work. Full stop. If you can find and keep exceptional people, do it. But here’s the honest truth: most hires aren’t great hires. A lot of folks don't want to discuss it, or want to challenge it. But deploying a great AI Agent is just easier, and less stressful, than making a mediocre hire.
16
3
37
4,236
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
Replying to @paulg
Reminds me of Shankar Vedantam’s line: “Those who travel with the current will always feel they are good swimmers; those who swim against the current may never realize they are better swimmers than they imagine.”
10
2
122
9,287
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
I think it must be a very interesting time to be in programming languages and formal methods because LLMs change the whole constraints landscape of software completely. Hints of this can already be seen, e.g. in the rising momentum behind porting C to Rust or the growing interest in upgrading legacy code bases in COBOL or etc. In particular, LLMs are *especially* good at translation compared to de-novo generation because 1) the original code base acts as a kind of highly detailed prompt, and 2) as a reference to write concrete tests with respect to. That said, even Rust is nowhere near optimal for LLMs as a target language. What kind of language is optimal? What concessions (if any) are still carved out for humans? Incredibly interesting new questions and opportunities. It feels likely that we'll end up re-writing large fractions of all software ever written many times over.
Shifting structures in a software world dominated by AI. Some first-order reflections (TL;DR at the end): Reducing software supply chains, the return of software monoliths – When rewriting code and understanding large foreign codebases becomes cheap, the incentive to rely on deep dependency trees collapses. Writing from scratch ¹ or extracting the relevant parts from another library is far easier when you can simply ask a code agent to handle it, rather than spending countless nights diving into an unfamiliar codebase. The reasons to reduce dependencies are compelling: a smaller attack surface for supply chain threats, smaller packaged software, improved performance, and faster boot times. By leveraging the tireless stamina of LLMs, the dream of coding an entire app from bare-metal considerations all the way up is becoming realistic. End of the Lindy effect – The Lindy effect holds that things which have been around for a long time are there for good reason and will likely continue to persist. It's related to Chesterton's fence: before removing something, you should first understand why it exists, which means removal always carries a cost. But in a world where software can be developed from first principles and understood by a tireless agent, this logic weakens. Older codebases can be explored at will; long-standing software can be replaced with far less friction. A codebase can be fully rewritten in a new language. ² Legacy software can be carefully studied and updated in situations where humans would have given up long ago. The catch: unknown unknowns remain unknown. The true extent of AI's impact will hinge on whether complete coverage of testing, edge cases, and formal verification is achievable. In an AI-dominated world, formal verification isn't optional—it's essential. The case for strongly typed languages – Historically, programming language adoption has been driven largely by human psychology and social dynamics. A language's success depended on a mix of factors: individual considerations like being easy to learn and simple to write correctly; community effects like how active and welcoming a community was, which in turn shaped how fast its ecosystem would grow; and fundamental properties like provable correctness, formal verification, and striking the right balance between dynamic and static checks—between the freedom to write anything and the discipline of guarding against edge cases and attacks. As the human factor diminishes, these dynamics will shift. Less dependence on human psychology will favor strongly typed, formally verifiable and/or high performance languages.³ These are often harder for humans to learn, but they're far better suited to LLMs, which thrive on formal verification and reinforcement learning environments. Expect this to reshape which languages dominate. Economic restructuring of open source – For decades, open-source communities have been built around humans finding connection through writing, learning, and using code together. In a world where most code is written—and perhaps more importantly, read—by machines, these incentives will start to break down.⁴ Communities of AIs building libraries and codebases together will likely emerge as a replacement, but such communities will lack the fundamentally human motivations that have driven open source until now. If the future of open-source development becomes largely devoid of humans, alignment of AI models won't just matter—it will be decisive. The future of new languages – Will AI agents face the same tradeoffs we do when developing or adopting new programming languages? Expressiveness vs. simplicity, safety vs. control, performance vs. abstraction, compile time vs. runtime, explicitness vs. conciseness. It's unclear that they will. In the long term, the reasons to create a new programming language will likely diverge significantly from the human-driven motivations of the past. There may well be an optimal programming language for LLMs—and there's no reason to assume it will resemble the ones humans have converged on. TL; DR: - Monoliths return – cheap rewriting kills dependency trees; smaller attack surface, better performance, bare-metal becomes realistic - Lindy effect weakens – legacy code loses its moat, but unknown unknowns persist; formal verification becomes essential - Strongly typed languages rise – human psychology mattered for adoption; now formal verification and RL environments favor types over ergonomics - Open source restructures – human connection drove the community; AI-written/read code breaks those incentives; alignment becomes decisive - New languages diverge – AI may not share our tradeoffs; optimal LLM programming languages may look nothing like what humans converged on ¹ xitter.cf/mntruell/status/201282… ² xitter.cf/anthropicai/status/201… ³ wesmckinney.com/blog/agent-e… ⁴ github.com/tailwindlabs/tail…
680
637
145
7,954
1,211,647
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
Replying to @tdinh_me
README files used to describe the software. now they ARE the software. that's the entire shift in one sentence.
1
2
117
Bryan Nelson 🇳🇱🇿🇦 🏊🚴🏃 retweeted
“Agentic engineering” Heard @steipete say this and it really hit home for me. Feels much more accurate to the way I’m building than “vibe coding”.
21
16
13
202
164,548