I agree in general, but think that titotal's specific use was fine. In my opinion, the main goal of that post was not to engage the AI 2037, which had already be done extensively in private but rather to communicate their views to the broader community. Titles in particular are extremely limited, many people only read the title, and titles are a key way people decide whether to eat on, and efficiency of communication is extremely important. The point they were trying to convey was these models that are treated as high status and prestigious should not be and I disagree that non-violent communication could have achieved a similar effect to that title (note, I don't particularly like how they framed the post, but I think this was perfectly reasonable from their perspective)

Thanks for this suggestion! I think it's helpful in a lot of contexts. But note that your response to point #1 assumes that the writer's aim must be to convince their interlocutor to change their mind. I disagree with this: public writing may have very different goals from those that a personal letter would have. A particularly important competing goal may be to rally a critical mass of readers to viscerally appreciate why the targeted view is wrong/bad. And this goal may be much better achieved by judgmental writing than by NVC (not least because bystanders are more apt to be interested enough to read a more emotionally engaging post featuring intellectual conflict).

NVC says: drop the adjectives and focus on indisputable facts, such as how you feel about the situation.

This seems like describing a totally different potential post. If titotal had written "why the AI 2027 model makes me sad" my reaction would have been 'why should I care about your emotional reaction?' - the important thing is whether the model is good. 

This strikes me as having some potentially adverse consequences. Although you suggest a possible extension to positive adjectives in the addendum, I am skeptical that this would be workable as a community norm. So I'll focus on the main proposal here.

I submit that it's generally undesirable in a truthseeking community to make negative evaluations more difficult or costly to express than positive ones. This is likely to tilt the field in favor of the latter. And there are already some nudges to skew positive -- both psychological (e.g., many people would rather avoid conflict) and structural (e.g., being positive is generally a better strategy in life for winning friends and influence). There are, of course, other social circumstances in which slanting the field toward positive feedback is desirable.

As others have implied, a Forum post is often intended to express a view about the nature of reality (~ a judgment) to third parties. To the extent there is "winning," the theory of change of such a post is that third parties update their views in a way that more closely tracks the way things are. That theory of change is harder to accomplish if one refrains from expressing a view about the nature of reality. And I don't think phrasing things as "I feel this model is badly flawed" would help things -- the reader understands that this is equivalent to a claim that the model is badly flawed.

That's not intended as a broader criticism of NVC, a topic on which I have no general opinion. But it does strike me as emphasizing ends like meeting participants' emotional needs and maintaining relationships rather than being focused on community truthseeking. I'm not someone who thinks that community norms should always maximize truthseeking over all other relevant considerations, but it is a rather important consideration (especially in the context of criticism of something with millions of page views, YouTube video views, etc.).

I agree that titotal should probably not have used the word "bad" but disagree with the reasoning.  The problem is that "bad" is extremely nonspecific; it doesn't tell the reader what titotal thinks is wrong with their models, even at a very low resolution.  They're just "bad".  There are other words that might have been informative, if used instead.

Of course, if he had multiple different kinds of issues with their models, he might have decided to omit adjectives from the title for brevity's sake and simply explained the issues in the post.  But if he thought the main issue was (for the sake of argument) that the models were misleading, then I think it would be fine for him to say that in the title.

I'm against most adjectives and also against most uses of "nonviolent communication", so let me offer a more fundamental reason to reduce your use of adjectives: It's just clearer writing/speaking to omit them. And the attempt to write more clearly probably forces you to think more clearly. Precisely-chosen adjectives are fine, but "bad" conveys more noise than signal. The fact that it may hurt someone's feelings is beside the point. You could just as easily offer a commentary on "AI 2027's excellent timeline models" and "excellent" would be similarly uninformative. OTOH, a bare description of the facts will hurt an emotionally intelligent person's feelings just as much as you calling it "bad", because they get the point. Depending on context, it might even hurt more, because they know it's a rationally-considered take and not an emotional one that you are likely to feel differently about once you calm down. 

The focus should be on speaking/writing in a way that will most clearly convey meaning. "The kitchen is dirty" obfuscates; "the dishes need to be washed" makes clear what you want (and conveys just as much "judgment" to all but the most obtuse). "Your presentation was too dry", OTOH, is better than "I didn't like it". But better still would be "I would like more examples of what these budget numbers mean for our day to day operations" or "expert consensus is that people will pay more attention if you sound excited when you give a presentation". 

To the extent that harshness is an EA norm, I think it's inherited from rationalist culture. In my experience with spaces like LessWrong, quite jarring critiques are fairly normal even for trivial things (e.g. “that argument is stupid”). There, bluntness is viewed as efficiency, getting bad ideas off the table faster. 

EA spaces are optimized for a different goal, and tone matters for that goal. We need people to feel welcomed, encouraged, and inspired to contribute; not like they’re auditioning for a role in a debate team. A good measure of how well we're doing on this is the fear people have of posting on the forum.

I haven’t read titotal’s post, so I won’t comment on that case, but I’ve definitely noticed the broader pattern Alfredo is pointing out. And I think we should be intentional about whether it serves the kind of community we want to build. 

Honestly I agree. Titotal’s work is fabulous (and I’m happy to use that adjective) but it’s usually not a good idea to put someone’s hackles up when you want to tell them they are wrong.

Judgements are also less informative. Knowing that titotal thinks a model is bad is not super useful except insofar as you want to defer to them. Eli's response was basically "yeah we agree with the facts you state but disagree that this makes the model bad," which feels like a clear illustration of the limitations of just saying "X is bad."

Besides voicing my general agreement with the OP, I'd like to bring a recent reflection by Scott Alexander (from https://www.astralcodexten.com/p/if-its-worth-your-time-to-lie-its), into this. 

"If one side [does X] to make all of their arguments sound 5% stronger, then over long enough it adds up. Unless they want to be left behind, the other side has to make all of their arguments 5% stronger too. Then there’s a new baseline - why not 10%? Why not 20%? This mechanism might sound theoretical when I describe it this way, but go to any space where corrections are discouraged, and you will see exactly this."

In the original post by Scott,"does X" is "lies". But I think the use of adjectives, specially when, as otherwise pointed, they are just generally used to indicate a general valence, and not a specific mistake, may be akin to this "small lies to make your arguments sound 5% stronger". And I furthermore think this is why it doesn't matter if the adjective is "bad" or "wonderful". 

Was titotal's post not a critique of 2027's models of the bad timeline? I had not interpreted it in the way you're describing.

Executive summary: This personal reflection advocates for reducing the use of judgmental adjectives—especially negative ones—on the Effective Altruism Forum, arguing that adopting a nonviolent communication (NVC) style centered on factual observations and personal feelings can prevent conflict, foster a more welcoming environment, and make critiques more effective without compromising clarity.

Key points:

1. Judgmental adjectives like “bad,” “boring,” or “stupid” are subjective and can provoke defensiveness, leading to unnecessary conflict and discouraging forum participation.
2. The author draws from nonviolent communication (NVC), emphasizing the value of describing personal feelings and observations rather than making evaluative claims.
3. Even when a judgment seems accurate or widely shared, its use may still alienate others and reduce cooperation or receptivity to feedback.
4. Adopting less judgmental phrasing—like “I disagree” or “I found it unengaging”—can preserve meaning while minimizing emotional harm.
5. The heuristic suggested is to avoid judgmental adjectives in high-stakes or emotionally charged contexts, while being more relaxed in casual ones.
6. Positive judgments can also be problematic in some contexts, as they still represent subjective evaluations that may distort communication or expectations.

This comment was auto-generated by the EA Forum Team. Feel free to point out issues with this summary by replying to the comment, and contact us if you have feedback.

In the example offered:
"Me: Ugh, the kitchen is dirty again. Why didn’t you do the dishes yesterday?"
What about the "Ugh" verbal expression of frustration and or disdain, as well as "Why didn't YOU do the dishes" which implies blame and or authority / superiority that 'you' have some right which another does not possess? (You have the right to dictate another person do dishes that you are apparently unwilling to do or take responsibility towards)

From my perspective of the information presented, this sounds more like spinning or twisting words around than resolving a root or core element of the problem. It is not that I disagree with the overall points you made in general about minding our emotions or the words we may choose to use, I do agree with that, however the part I don't agree with is the fixation on adjectives with exclusivity. 

"2. But the adjective I chose conveys something specific that I can’t otherwise convey by just stating my feelings. Saying “I didn’t like your presentation” is just not the same as “Your presentation was too dry”."

The difference is subjective, quire minor semantics of degree in which is stated that 'the presentation was not liked'. saying 'the presentation was too dry' is still an insult and arguably even worse if you knew it was very much not a compliment.

Communication is not about "winning", it should be about collaboration, not competition.
Communication is not "violence". Violence is self-defined as an act of violation. Truth cannot be a violator. What you'd most likely mean is that untruth, lies, psychological abuse, censorship, dishonest communication tactics, those are technically real manifestations of words being used to stifle and prevent true communication and embolden injustices in some manner. But they are not acts of communication turned violent, but acts of violence upon and through the concept of "communication" as a proxy.