Reading Advantage delivered a suite of educational apps for your school designed to help bilingual students improve both English and subject performance.
I’m so sick of Markdown right now.
It’s amazing for humans and LLMs, but impossible to process or manipulate otherwise.
Is there anything that renders like Markdown (for humans) but has a strict structure (like XML/YAML/etc) so it’s easy to manipulate?
Ideally, a superset of Markdown, so it “just works” and renders like normal Markdown. But with added structure that makes it easy to parse and verify and manipulate.
I’m building my own thing currently but I’d rather not.
I probably use Chinese models every day, so I know their limitations.
The "fashion" arose from people touring bench-maxed models being sold as "just as good" when they weren't close. DeepSWE was a much better indicator of what you would see in real use.
But I'm glad 5.2 improved. I hope every other model I use keeps improving.
GLM kicks a** on DeepSWE
Recently it had become fashionable to beat Chinese OSS models on new benchmarks like DeepSWE. I have been arguing there is nothing fundamentally wrong with Chinese labs, and it is matter of only time (and generation of models). I stand vindicated.
GLM-5.2 shuts the mouth of all the critics by beating Claude Opus 4.6 released in Feb 2026. The gap is barely 3-4 months.
So apparently Google ships a Gemini Nano 4B LLM (context limit : 9216 tokens) baked into chrome
I tried to expose it out as an OpenAI-compatible API for my local, so basically no API key, no external network calls, no need of ollama
Demo and repo below ↓
Among the worst damage that the AI hype train has delivered to itself in the past years is creating the illusion that there is no value in a computer science degree.
To be fair, a lot of people thought that CS = coding, which is totally wrong. And CS programs absolutely need to be reformed and adjusted to account for advances in AI.
But constantly ranting about AI replacing software engineers and destroying all jobs was absolutely not helpful either.
Another own goal for the industry.
I don't think that is clear at all. I have subscriptions for all SOTA models and for Kimi, Minimax, Xiaomi, Doubao, and GLM, which I spend multiple times as much on as the SOTA models, so I know ALL about them.
China has some interesting architectures, but they're not close YET.
My daughter spent 4 years earning her bachelor’s degree, followed by another year doing research at UC San Diego and after scoring high on the MCAT to get accepted into medical school. After 4 more years of med school, she’ll be competing for residency — a total of 10 years of hard work and massive debt.
Meanwhile, she’ll be up against foreign doctors who only spent 6 years in their home countries to become physicians.
How is this fair to American students who invest a decade of their lives and hundreds of thousands in debt to become doctors? Why aren’t we putting American doctors first? Who’s going to compensate our kids for this burden?
America First means taking care of our own. 🇺🇸
Seventy years ago, rural farmers got health care despite having no insurance and next to no money. What happened?
Medicine for insanely now expensive, better, largely sure to technology.
But there's another factor -- regulation. Reduce regulation on drugs, hospitals, and practitioners.
The two best multi-agent plans:
Global: OpenCode Go ($10 but needs a bigger plan) https://t.co/GreEtp57lq
China mainland: ByteDance Ark Coding Plan 200 RMB = $30 -- perfect size) https://t.co/WJDXKvqJAJ
Social Security wouldn't need to run out of it were ... you know ... social security -- a safety net. Just make all future benefits equal to the poverty line. Then set withholding to the necessary amount to cover the estimated budget. No calls. Simple math.
Social Security wouldn't need to run out if we eliminated it for billionaires, required millionaires to pay more into it, and stopped taxing it. You can thank Ronald Reagan for this damage, btw.😡
i have to say "a single person having a trillion dollars does nothing to hurt poor people or worsen global poverty" is a very stupid thing to say about *elon musk* a guy who quite famously has already used his billions to get the us government to end all international aid
@conor64 I'm not one who worries admit such things because when you look at this historically, individuals have more influence than they have for most of US history. Universal suffrage. Primaries. "Independent journalists." Before PACs, there were still unions and other organizations.
I bought my last vehicle cash.
They still had me come back twice, sit in with the finance manager for 45 minutes filling out paperwork that I specifically asked to already be filled out so it wouldn’t take 45 minutes.
Dude kept trying to sell me on stuff.
Dealerships cannot die fast enough.
Agreed.
Three years ago: Reading Advantage used gpt-3.5 API. I used chat.
One year ago: I used Claude Code API for my first real end-to-end published project.
Ten months ago: Subscriptions saved the world.
Now: seven subscriptions, OpenCode, multi-agent loops.
Two years ago, reasoning models did not exist. GPT-4o and Sonnet 3.5 were SOTA.
One year ago, the best publicly available model was o3, and 99% of those who used it (i.e., 99% of 7% of OpenAI customers) were using it as a chatbot.
Today (i.e., right now as I type this), I have Codex building several pieces of personalized software and doing a large-scale data management project for me. The data management project has been running autonomously for nearly 4 days. I have multiple working apps on my phone that I use daily and which were built entirely by Codex.
Projecting this trend to June 2027 is absolutely mind-boggling. The world will change.
A lot of people (and algorithms) assume I support the right and Republicans. I don't. I just oppose Democrats.
This is the party that fights against:
* Protecting voters and rights
* Eliminating government fraud
* Following the law
* Telling the truth about their motives