Bram Zijlstra

Every now and then I read posts about 90% of coding will be done by AI this year and I can't lie it makes me a bit nervous. Then a study likes this drops and I can relax a bit

You’ve probably heard about how AI/LLMs can solve Math Olympiad problems ( deepmind.google/discover/blo... ). So naturally, some people put it to the test — hours after the 2025 US Math Olympiad problems were released. The result: They all sucked!

There are plenty of cases where survivorship bias doesn’t apply. We just don’t remember them.

Do you think vibe coding is a gradual or fundamental change in technology?

Errors with a 200 status code is like gift wrapping a turd

An infinite amount of monkeys with an LLM can write the entire the complete works of Shakespeare

My doctor told me he's into vibe surgeries lately

I think the reason why LLMs are overconfident is because we keep saying telling "You are an expert in" literally anything

"Why flake8? I use Ruff"

Electron apps are the Monkey's Paw to someone wishing cheaper RAM

Anyone know best practices / tips for improving LLM quality on classifying long texts? For shorter input few-shot learning and finding the best examples work well, but this not very practical with longer texts. #databs

We'll have AI alignment before MS Word alignment

An agent is an LLM that uses tools. A tool is someone who keeps saying '2025 is the year of the agents'

Today was a day where all emails found me well

Interesting benchmark from Adyen on multi-step reasoning. New benchmarks are great in establishing a baseline for historic models, all new models should be treated with suspicion. #databs huggingface.co/blog/dabstep

Really want to like cursor but I am completely Pycharm brained it seems. Anyone else have the same ? Which copilot did you go for?

“Our company has moat” The moat:

DeepSeek is not refusing the Tiananmen Square question, though it gets the answer wrong.

So, look. I'm sure I'm in the minority here on Bluesky in believing that training AI systems isn't copyright infringement. But, also. Dude. There's no way OpenAI can make this argument without looking very, very silly.

Will copilots (eventually) negatively impact open source library development? Major updates will break any copilot, so it disincentives doing so.

Baffling to see the disclosure around DeepSeek Last I checked the tech industry, we celebrate small teams pushing the industry forward, coming up w novel ways to build software more efficiently. And sharing it. Yet now there's a class of folks who think this is some bad thing?

If I would be a dictator for a day I would ban the word 'syncing' outside the computer domain.

DeepSeek released a whole family of inference-scaling / "reasoning" models today, including distilled variants based on Llama and Qwen Here are my notes on the new models, plus how I ran DeepSeek-R1-Distill-Llama-8B on my Mac using Ollama and LLM simonwillison.net/2025/Jan/20/...

Lol, if only we could give R1 a hug.

I am working with Deepseek R1 today and used it for a simple classifier with few-shot learning. Looking at the thinking tokens I think I know now how anxiety and extreme perfectionism looks like

When you replace cheese on your pizza with feces you can't claim to save money on ingredients, you just make everything taste like shit.

Anyone else feel really proud when o1 needs to think a long time for your prompt?

Google should create a "no captcha" game where you have to pretend to be a robot and get points for tricking the algorithm

Porques no los neither?

If your AI boyfriend is sharing weights with other women, is that considered cheating?

Do other things have a 'Bitter Lesson' like Machine Learning? Restaurants: It's always the salt and butter that makes it better Stress: It's finishing the task you dread most that will reduce your anxiety

“If you could walk a mile in my shoes” is the Docker equivalent of life experiences

My running coach says I’m not a real runner until I fix my form. I think he's gait keeping

The world is awful. The world is much better. The world can be much better. All three statements are true at the same time. Understanding this is key to solving big global problems. We believe data & research can help us understand both the problems we face & the progress that’s possible. 🧵

When the chips in small size Make the compute go twice, that's a Moore #databs

my hot take: data IS NOT better than pure vibes because it has a way of causing people’s brains to shut off. i think a hallmark trait of the currently growing generation alpha will be their skepticism, questioning everything they see this is the way. data is not enough

I've been tracking my stress levels on Garmin and a stupid tip is just to stress less. It works 🤷‍♂️

Tired: Saying "It's so over" after every AI breakthrough Wired: Saying "It's so over" in normal conversations every time you hear something new

I thought buying an M2 would set me for years but seeing people using GPT-4 like models on M4's make me feel like I'm programming on a wet cardboard.

Normalise interpersonal consensus forming on what is socially acceptable or not

trending topics is a great way to break out of my echo chamber man things seem bleak outside my echo chamber

I once took a course in Heideggers "Seind Un Zeit" and I kept changing my mind between "this is genius and revolutionary" and "this is completely incomprehensible", often multiple times per page and without a definitive conclusion

you can spot a hobby project on the internet because it uses 100% IPv6

The correlation between how much I like my post and how well it is doing is not a positive number

Affirmations are prompt instructions for humans

2005: You can't trust anything that's on Wikipedia 2025: You can't trust anything besides Wikipedia