A new way to reflect on how you use Claude | Hacker News Reader

A new way to reflect on how you use Claude

anthropic.com

44 pointssurprisetalk58 comments

> If you can't generate a report, it may be because you don’t have Memory turned on.

Has anyone found success with memory and Claude Code? I found Opus's 4.8 memories largely lacking value. Memories are too verbose/specific/yet-generic for the value captured.

Instead, I've been holding "retros" with the agent immediately after a session, and _those_ responses have been typically fantastic. It has ideas for coding changes, spots unmentioned small bugs, suggests invariants, principles to adopt, lint rules, tooling tweaks, skill-file updates, follow up work, all kinds of stuff.

Memories have mostly been detrimental to me, I have disabled it completely after many moments of frustration.

In general Claude would decide to save irrelevant memories, or apply a memory that was lifted during work on a project to a completely different context where it wasn't relevant, or saved a memory from how to use a tool while I was experimenting with the toll and silently apply it breaking my workflow.

I can't remember a single instance where it was helpful so I just disabled it to not have to deal with yet another cognitive overload.

> I found Opus's 4.8 memories largely lacking value.

Similar thoughts, I disabled memory for the web UI.

> I've been holding "retros" with the agent immediately after a session, and _those_ responses have been typically fantastic.

I used a skill instead to update a ai docs directory as needed.

In any case, I'm moving coding tasks to Codex, Claude will be demoted to frontend design only. I will not entertain the idea of API billing/extra usage or a 2nd Max account once the extra 50% Claude Code usage limits ends on 13 July [1].

[1]: https://news.ycombinator.com/item?id=48126429

Self-directed memories have been awful.

That said, I have had decent success in self-feedback-loop skills. I encapsulate a long process into a skill and the last step is to make changes as needed to the skill itself for things it ran into. I think the isolation of "this specific thing" is what makes it higher signal. Project or god-forbid global memory is just a recipe for hallucinations in my experience.

For me, the memory capability is a holding zone for deviations from behavior I'm looking for. I've found that a "retro" after to process "memories" and create CLAUDE.md instructions from them instead works more reliably longer term. I don't do it every session. I've got a separate tool which tracks memories on a per-project basis and recommends when to hold that retro based on accumulated memories.

Seem to work fine for me but I usually add 1 sentence things to keep the agent from long winded rediscovery.

Remember credentials to query database are in secret store x

Remember company documentation is in platform y

Remember bug fixes always need tests added

> In our interviews with users, a common theme that’s emerged is a desire to better understand how, exactly, AI can be integrated into daily life.

I don't want to integrate AI into daily life.

Meh, I have it integrated as a replacement for search engines at this point.

I don’t even know that Claude is inherently better or if it’s more the lack of ads.

I did this but have (partly) reverted. I don't always want to read a wall of text that an AI regurgitates for search. The google AI snippet (1 short para) does seem better than the typical ChatGPT response.

I'd argue that it's not that Claude is better, but that Google has gotten worse.

Search engines without ads exist.

Try noai.duckduckgo.com. Decent search results, actual sources, no risk of hallucination, no extreme energy cost.

Kagi is an even better option, IME.

I don't know why this sentiment isn't more common. LLMs are basically two things: 1. chatbots 2. automation for the process of googling something and copy-pasting the first result. The only people I know who use it for more than those two things ask them to perform tasks they can't do reliably, don't check the output, and then they're just wrong most of the time. If they had just used it as slightly different google search they'd have been fine.

When I have a good idea that I can hit what I want in a single query, I still use search engines directly a lot. But I have found AI is pretty decent at automating the process of making a vague search, picking up from the search better search terms, then making a couple of other wrong searches, then refining down to what I really want and potentially pulling together an answer from multiple sources. I know how to do all that, but the AI can do in a minute what may take me 15 minutes.

I'd say I also know I can still do it, but... as the search engines deteriorate it is getting somewhat harder to do this by hand than it used to be. I still do this by hand sometimes for cases where I want the exploration of a topic for myself, rather than a focused answer where I don't really care about what I learn along the way, and it's getting harder. I don't know that it'll converge at "zero value" but the search result pages seem like they're just... harder to use for this than they used to be, though it's hard to put my finger on how.

Then…don’t? You are free to not do so, while others (like myself) can choose the opposite.

Tell that to our employers

Using AI for specific tasks at work is definitionally not "integrating AI into your life."

You might use a badge to open a door at work but that doesn't mean your integrating badge-door access control systems into your daily life. It's a tool that you only use at work.

To be fair, the article doesn’t quite deliver on this promise. The examples are mostly focused on improving work-related workflows, so I guess that’s what they think “daily life” is.

To most of the people working on these AI systems I'm willing to bet work in 90% of their waking life or more.

If someone is doing something like that for 90%+ of their waking life it means it's not work; it's an enjoyable pastime for which they're being paid well.

Then this isn’t for you. That’s fine.

>how, exactly, AI can be integrated into daily life.

Aka "what is it good for?"

"Cool product you have there... but what is it for?"

I just feel that any attempt of a service I use to summarize and analyze my interactions with it, whether it's the AI tool usage patterns or the music I listened to the most over the past year, makes me feel creepy and makes me want to use the service less. Imagine if your local grocery store came back to you saying that you ate this many chocolate bars over the year. Thanks, I know that you know that, but I don't want you to show me that you know that.

I feel this way too except for the music one. The whole reason I use spotify is because I want their algorithm to figure out what music I like.

> ... figure out what music I like.

Isn't that something you determine yourself? I listen to a lot of heavy metal because I like it. I also wander around bandcamp and find stuff that isn't heavy metal, but enjoyable, so I buy it. My brain is my algorithm.

I do determine it for myself but sometimes I just want to hit play and hear new stuff.

not everyone has the time to trawl through thousands of bands to find music they like

Some people enjoy listening to music, not looking for music.

I would love it if a grocery store, or Uber, or Amazon, made it easy to see my yearly spend. Instead them to intentionally obscure this and try to prevent you from seeing a big number.

https://www.youtube.com/watch?v=Hx7Vv5pqpHg

In claude code, there's a built-in skill `/insights` which gives you a report on how you've been using claude code and where you could improve.

Is it any good?

No, it’s superficial slop

Speaking as a big proponent of Claude code in general, which I find to be revolutionary and useful - there is no value in that report. To be honest, the people who I know who like that report are the ones who are getting sycophantically gaslit by the models more than they should.

> which I find to be revolutionary and useful - there is no value in that report

:) I went through some Claude documentation that apparently came with my job's paid subscription.

Besides some vague description of how to use the API, it was just fluff. For example there were exactly zero hints on how to do your prompts.

How people prompt and the output that ensues is valuable data. So they would probably want variety as a free live experiment.

So a paid product that comes with no documentation, basically.

Not an anthropic fan but in fairness people know how to use it in the most literal sense. Few tools come with documentation on how to use them "well"

agreed, but IMO thats been true for most off-the-shelf skills I've seen. Instead I built my own "/meta" skill to improve my workflows, which I can use in targeted ways across my projects

This product feels bad and sloppy, so I’ll give my hypothesis for why this was built:

At this point, Anthropic is likely having Claude itself propose and build features autonomously based on providing it with raw user feedback. This could be one example. Which is why it has an eerie sense of redundancy and pointlessness (“You mostly used Claude to automate work and home tasks”, etc.).

Interesting, I am not able to see it, and I have memory enabled.

Same. Not sure if they are doing a gradual rollout or what's up.

I've integrated AI into my daily life by installing the Gemini voice app onto my phone, which I typically may use once a week, and adding Gemini and Claude bookmarks to my browser which I use all the time - but mostly Gemini since free usage is effectively unlimited.

That's all the integration I need. I don't need OpenClaw running 24x7 trying to hack it's way into my gmail.

I think there's a germ of a good idea here, but really this needs to be data that is presented in a way that encourages human thought and interpretation and not something claude predigests and interprets _for_ you. I found the whole tone as it's executed incredibly off-putting and cringe inducing.

How would you present a collection of your past text conversations that it encourages human thought and interpretation?

> It lets you easily track and visualize how you use Claude, and decide whether that time aligns with your goals.

i haven't yet encountered/achieved ai summarization technique nuanced enough to be called reflection and this seems basic too with no mention of behavioural patterns or workflow suitability

It seems saying "Your way to use Claude is wrong somehow, you should improve the way to use AI", instead of "Our tools have some wrong aspects and we will do our best to improve them".

It's actually quite normal to try to acquire skill at using tools. It's very normal for toolmakers to share insights and expertise in how best to use the tools they sell.

For most people, I guess it will say you need to take an English grammer course and draft your own email.

Why would a product talk its users out of using it?

Remember that time grok told people not to use it for second opinions on their medical records? ...on second thought that might be an exception that proves the rule.

> you need to take an English grammer course

And a spolling one too.

It is inevitable that a post criticizing grammar or spelling will, itself, contain a grammar or spilling error.

This looks like mindless data tourism to me. I don't see much distinction between this and something like Spotify's annual "wrapped" feature. The information theoretical equivalent of junk food.

This reads very promotional, like an ad.

Is anthropic paying you guys for all this free advertising dang? This is pathetic.

checked everything - still don't have this function. any clue how to get it?

Here's my take as someone who was using Claude everyday for every little thing and now I've deleted my account because I didn't like what it was doing to my mind.

What goes totally unmentioned in the article is that this feature is designed to help mitigate dark usage patterns. My major concern with chatbots is excessive usage can lead to AI psychosis or negative rumination (depression). A feature that makes user's more self-aware of their usage patterns is a good thing. Making the user more self-aware is a necessary first step that will precede any kind of intervention to reduce their reliance.

Where this fails is it frames the intervention as a moderation problem. It may seem counterintuitive, but moderation takes more self-control than elimination. If you struggle with your relationship to LLMs, then every time you make a choice whether or not to engage with a LLM is an opportunity to struggle. The more you struggle, the worse you feel.

Obviously Anthropic cannot advocate for churning from their product. The psychological stickiness of their product is its primary selling point for investors. When they say "set quiet hours and breaks" it frames this as a user problem. Just get good bruh, it's not that the technology hallucinates or is sycophantic, or basically designed to be a AI girlfriend / boyfriend, it's a skill issue. Rather than a technology being applied incorrectly and a company floundering to hook users before they try to jack up the prices to stay solvent.

I find the "AI Fluency" course particularly ridiculous.

> Build AI skills that support your original thinking

This is a straight-up lie. When you outsource your thinking to Claude then your ability to produce original thought degrades. The whole framework for using LLMs is the sort of thing you see from tech bros on Twitter trying to sell online courses. It reminds me of the intellectual yet idiot essay by Nassim Taleb[1]. Don't let Anthropic tell you how to think under the guise of doom trolling[2] and tech bro "thinking frameworks". Think for yourself!

[1] https://nassimtaleb.org/2016/09/intellectual-yet-idiot/

[2] https://www.nytimes.com/2026/06/17/opinion/ai-dangerous-open...

Reply on news.ycombinator.com