Posted by simonw 7 days ago
People posting their subjective experience is precisely what a lot of these pieces should be doing, good or bad, their experience is the data they have to contribute.
The plural of anecdote is not data. These subjective posts about experiences vibe coding, etc. may be entertaining but if you read 10 of them it doesn't give you an objective view of the state of LLMs. It gives you 10 opinions by 10 people who chose to blog about how they felt using a tool.
Second of all, Simon's content are often informative, more or less sticking to the facts, not flame bait. I never upvote or flag any content from anyone.
I called out the terrible scatter plot of the latitude/longitude points because it helped show that this thing has its own flaws.
I know so many people who are convinced that ChatGPT's search feature is entirely useless. This post is mainly for them.
Those are the kinds of things I look out for and try to write about.
I didn’t feel that he was framing it as _revolutionary_ it felt more evolutionary.
Simon, for every person miffed about your writing, there is another person like me today who said “ok, I guess I should sign up for Simon’s newsletter.” Keep it up.
It’s easy to be a hater on da internet.
42lux, if you have better articles on AI progress do please link them so we can all benefit.
I wanna know when my research goblin can run on my box with 2x 3090s.
I skipped half the article to get to the point, went back and re-read and didn't miss much.
This is fine. He is his own person and can write about whatever he wants and work with whoever he wants, but the days when I'd eagerly read his blog to get a finger of the pulse of all of the main developments in the main labs/models has passed, as he seems to only really cover OpenAI these days, and major events from non-OpenAI labs/models don't seem to even get a mention even if they're huge (e.g. nano banana).
That's fine. It's his blog. He can do what he wants. But to me personally he feels like an OpenAI mouthpiece now. But that's just my opinion.
My most recent posts:
- https://simonwillison.net/2025/Sep/7/ai-mode/ - Google/Gemini
- https://simonwillison.net/2025/Sep/6/research-goblin/ - OpenAI/GPT-5
- https://simonwillison.net/2025/Sep/6/kimi-k2-instruct-0905/ - Moonshot/Kimi/Groq
- https://simonwillison.net/2025/Sep/6/anthropic-settlement/ - Anthropic (legal settlement)
- https://simonwillison.net/2025/Sep/4/embedding-gemma/ - Google/Gemma
So far in 2025: 106 posts tagged OpenAI, 78 tagged Claude, 58 tagged Gemini, 55 tagged ai-in-china (which includes DeepSeek and Qwen and suchlike.)
I think I'm balancing the vendors pretty well, personally. I'm particularly proud of my coverage of significant model releases - this tag has 140 posts now! https://simonwillison.net/tags/llm-release/
OpenAI did get a lot of attention from me over the last six weeks thanks to the combination of gpt-oss and GPT-5.
I do regret not having written about Nano Banana yet, I've been trying to find a good angle on it that hasn't already been covered to death.
You are. Pretty much my main source these days to get a filtered down, generalist/pragmatic view on use of LLMs in software dev. I'm stumped as to what the person above you is talking about.
OT: maybe I missed this but is the Substack new and any reason (besides visibility) you're launching newsletters there vs. on your wonderful site? :)
I wrote about how it works here: https://simonwillison.net/2023/Apr/4/substack-observable/
Sometimes I feel that there's pressure for me to be a full blown newspaper covering everything that happens in a multi-billion dollar industry.
I wrote about my approach to that here: https://simonwillison.net/2024/Dec/22/link-blog/#trying-to-a...
> he feels like an OpenAI mouthpiece now
That seems a little harsh. But, I felt the same about older blogs I used to read such as CodingHorror. They just aren’t for me anymore after diverging into other topics.
I really liked this article and the coining of the term “Research Goblin”. That is how I use it too sometimes. Which is also how I used to use Google.
https://news.ycombinator.com/submitted?id=simonw
Or take a look at his website:
At least you admit it's your opinion. Maybe that's your bias showing?
Personally I generally enjoy the blog and the writing, but not so much this post. It has a very clickbaity title for some results which aren't particularly impressive.
I find it informative that search works so well. I knew it works well, but this feels like step above whatever Gemini can do, which is my go to workhorse for chatbots.
This is on purpose, because we want good stories to get multiple chances at getting noticed. Otherwise there's too much randomness in what gets traction.
Plenty of past explanations here:
https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...
The 8 hours seem not to count if you submit under a different domain or do they reset after each try?
Would also be great if you would answer emails especially if they are related to GDPR. You have two of them in your inbox from over 6 months ago send from the email in my account.
From https://en.wikipedia.org/wiki/Social_news_website : "A social news website is a website that features user-posted stories. Such stories are ranked based on popularity, as voted on by other users of the site or by website administrators."
The article was recently published, users on HN submitted the article. Other users thought it interesting and upvoted. Earth has different time zones (I understand it's difficult for americans to grasp) and so different people are active at different times.
I'd say trust is a pretty reasonable way to assign attention.
I guess the fairest way might theoretically be to require everything to be submitted anonymously, with maybe authorship (maybe submissionship) only being revealed after some assigned period?
This is better for the incubants, but would require a huge amount of energy compared to "Oh, simon finds this interesting, I'll take a looksy".
The AI space is full of BS and grift, which makes reputation and the resulting trust built on that reputation important. I think the popularity of certain authors has as much to do with trust as anything else.
If I see one of Simon’s posts, I know there’s a good chance it’s more signal than noise, and I know how to contextualize what he’s saying based on his past work. This is far more difficult with a random “better” article from someone I don’t know.
People tend to post what they follow, and I don’t think it’s lazy to follow the known voices in the field who have proven not to be grifting hype people.
I do think this has some potential negatives, i.e. sure, there might be “much better” content that doesn’t get highlighted. But if the person writing that better content keeps doing so consistently, chances are they’ll eventually find their audience, and maybe it’ll make its way here.
Saying that someone ought to write better consistently for them to "make its way here" leans completely into the cult of personality.
I think following people would be better served though personal RSS feeds, and letting content rise based on its merit ought to be an HN goal. How that can be achieved, I don't know. What I am saying is that the potential negatives are far far understated than they ought to be.
> Saying that someone ought to write better
I did not say someone ought to write better. I described what I believed the dynamic is.
> I think following people would be better served though personal RSS feeds
My point was that this is exactly what people are doing, and that people tend to post content here from the people they follow.
> letting content rise based on its merit ought to be an HN goal
My point was that merit is earned, and people tend to attach weight to certain voices who have already earned it.
Don’t get me wrong. I’m not saying there are no downsides, and I said as much in the original comment.
HN regularly upvotes obscure content from people who are certainly not the center of a cult of personality. I was attempting to explain why I think this is more prevalent with AI and why I think that’s understandable in a landscape filled with slop.
Your Exeter cavern quandary was not exactly sorted. https://simonwillison.net/2025/Sep/6/research-goblin/#histor...
They are quite old and very well documented, so how on earth could a LLM fuck up unless, a LLM is some sort of next token guesser ...
I made fun of its attempt at drawing a useless scatter chart.
That example wasn't meant to illustrate that it's flawless - just that it's interesting and useful, even when it doesn't get to the ideal answer.