Performance and telemetry analysis of Trae IDE, ByteDance's VSCode fork

See the authors response. He or she says it doesn’t matter either way

https://news.ycombinator.com/item?id=44706580

It's also a good alternative to Obisdian if you don't need smartphone support.

Thank you! The timing of this comment is perfect

Eclipse (as in ecosystem) is fairly popular in Enterprise, but since it exposes all the knobs, and is a bona fide IDE which has some learning curve, people stay away from it.

Also it used to be kinda heavy, but it became lighter because of Moore's law and good code management practices all over the board.

I'm planning to deploy Theia in its web based form if possible, but still didn't have the time to tinker with that one.

Installing the VSCode extension pack for Java runs a headless version of Eclipse JDT under the hood, which isn’t quite what I think of as lightweight.

This seems perfectly rational. If you're already entrusting all your technical infrastructure to AWS, then adding another AWS service doesn't add any additional supply-chain risk, whereas adding something from another vendor does do that.

Yeah but this is industry standard

not saying this is good but everyone do this

It can be disabled btw. And no telemetry is collected on first run.

I keep it disabled for both Dart and Flutter.

As a somewhat paranoid person I find this level of paranoia beyond me. Like do you own a car? Or a phone? A credit card? Walk around in public where there's cameras on every block? I don't agree with it at all but the world we're living it makes it impossible to not be tracked with way more than (usually anonymized) telemetry data.

Or "I'm a power user" of sorts. Probably a very small minority of users fiddle that setting.

Dang said a similarly small minority of users here do all the commenting.

I suspect many (or all) VPNs probably do secret logging. People will do their most interesting secret activities on those.

almost as if it was a trap... ;p

Hate to break it to you, but /etc/hosts only works for apps that use getaddrinfo or similar APIs. Anything that does its own DNS resolution, which coincidentally includes anything Chromium-based, is free to ignore your hosts file.

When the nefarious actor is already inside the house, who knows to what lengths they will go to circumvent the protections? External network blocker is more straightforward (packets go in, packets go out), so easier to ensure that there is nothing funny happening.

On Apple devices, first-party applications get to circumvent LittleSnitch-like filtering. Presumably harder to hide this kind of activity on Linux, but then you need to have the expertise to be aware of the gaps. Docker still punches through your firewall configuration.

Set up your router to offer DNS through pihole and everything in your network now has tracking and ads blocked, even the wifi dishwasher.

So that these domains are automatically blocked on all devices on a local network. Also, you can't really edit the hosts file on Android or iOS, but I guess mobile OSes are not part of the discussion here.

Although there are caveats -- if an app decides to use its own DNS server, sometimes secure DNS, you are still out of luck. I just recently discovered that Android webview may bypass whatever DNS your Wi-Fi points to.

If you have multiple devices on the same LAN, all of them will use the pihole.

The hosts file doesn't let you properly block domains. It only lets you resolve them to something else. It's the wrong tool for the job.

Meta is pretty much number one, Google is pretty much number two. Whoever number three is, they are very far behind.

For what it's worth, I do use Google products personally. But I won't go near Facebook, WhatsApp, or Instagram.

Yes.

And the other side of the middle ground, Grafana being AGPL but requiring you to disable 4 analytics flags, 1 gravatar flag, and (I think) one of their default dashboards was also fetching news from a Grafana URL.

https://github.com/grafana/tempo/discussions/5001#discussion...

(Yes, that's for Grafana tempo, but the issue in `grafana/grafana` was just marked as duplicate of this.)

> Yes, why do people use products from Microsoft, Apple, Google, Amazon, ...

I work at Apple, so I’m not concerned about being monitored—it’s all company-owned equipment and data anyway.

It was the same when I worked at Microsoft. I used Microsoft products exclusively, regardless of any potential privacy concerns.

Employees at Google and Amazon do the same. It’s known as “dogfooding”—using your own products to test and improve them (https://en.wikipedia.org/wiki/Eating_your_own_dog_food).

As for why people outside these companies use their products, it usually comes down to two reasons: a) Their employer has purchased licenses and wants employees to use them, either for compliance or to get value from the investment; or b) They genuinely like the product—whether it’s because of its features, price, performance, support, or overall experience.

Because the alternatives suck.

In this case, the software being analyzed is the alternative that sucks.

It is literally spying on the user.

Unless you're somehow saying telemetry doesn't report anything about what a user is doing to it's home server.

Anonymized or not, opt-out telemetry is plain spying. Go was about to find out, and they backed out the last millisecond and converted to opt-in, for example.

Any monitoring of my system without my explicit permission is spying.

> We need to bring back shaming and social isolation

HN is not your army, and it isn’t a theater of ideological battle.

Please don’t do that here.

Or anything else from their party..

The mere fact that disabling telemetry does not at all disable telemetry is enough for it to be called spyware.

Have Bytedance produced literally anything to that assumption unreasonable?

Emacs is great, sure, but it lacks a decent text editor.

So does vi.

VSCodium I think was abandoned. It was extremely buggy last time I used it.

Either way it uses electron. Which I hate so much.

> Microsoft is content with funding it, the price is your telemetry (for now).

The price of VSCode is halo effect for Azure products

I thought with VS Code the price is that it entices you into using Azure, where the enterprise big bucks are made.

  > VIM is a text editor, not an IDE

What is there in an IDE today, that is missing from (n)vim? With the advent of DAP and LSP servers, I can't find anything that I would use a "proper" IDE for.

And electron is standard tooling? LOL

  > I install vscode from scratch, install a few extensions I need, set 3 or 4 settings I use regularly, and bang in 5 minutes I have a customized, working environment catered for almost any language.

Hell, I will even feel comfortable in a vi terminal, though that's extremely rare to actually find. Usually vi is just remapped to vim

Edit:

The git folder with *all* my dotfiles (which includes all my notes) is just 3M, so I can take it anywhere. If I install all the plugins and if I install all the vim plugins I currently have (which some are old and I don't use) the total is ~100M. So...

I'm sure you didn't mean to, but you've crossed into being aggressive with another user. Please don't do that on HN—not with anyone, and least of all new users who deserve to be welcomed and treated charitably, not harrassed for not already knowing HN's arcane and rather primitive formatting rules*.

https://news.ycombinator.com/newsguidelines.html

* (which, btw, are at https://news.ycombinator.com/formatdoc)

I think we can cut the guy some slack for using machine translation given that English isn't his native language.

I bet I sound a lot dumber than this when I'm renting servers in Hong Kong in Cantonese.

Edited to add space between the enumerated list.

Hi HN, I was evaluating IDEs for a personal project and decided to test Trae, ByteDance's fork of VSCode. I immediately noticed some significant performance and privacy issues that I felt were worth sharing. I've written up a full analysis with screenshots, network logs, and data payloads in the linked post. Here are the key findings:

1. Extreme Resource Consumption: Out of the box, Trae used 6.3x more RAM (~5.7 GB) and spawned 3.7x more processes (33 total) than a standard VSCode setup with the same project open. The team has since made improvements, but it's still significantly heavier.

2. Telemetry Opt-Out Doesn't Work (It Makes It Worse): I found Trae was constantly sending data to ByteDance servers (byteoversea.com). I went into the settings and disabled all telemetry. To my surprise, this didn't stop the traffic. In fact, it increased the frequency of batch data collection. The telemetry "off" switch appears to be purely cosmetic.

3. What's Being Sent: Even with telemetry "disabled," Trae sends detailed payloads including: Hardware specs (CPU, memory, etc.) Persistent user, device, and machine IDs OS version, app language, user name Granular usage data like time-on-ide, window focus state, and active file types.

4. Community Censorship: When I tried to discuss these findings on their official Discord, my posts were deleted and my account was muted for 7 days. It seems words like "track" trigger an automated gag rule, which prevents any real discussion about privacy.

I believe developers should be aware of this behavior. The combination of resource drain, non-functional privacy settings, and censorship of technical feedback is a major red flag. The full, detailed analysis with all the evidence (process lists, Fiddler captures, JSON payloads, and screenshots of the Discord moderation) is available at the link. Happy to answer any questions.

or maybe hn shouldn't have special formatting and support markdown

I don't think the post you're replying to deserves what feels like an aggressive dressing-down. The reality is that post formatting on HN is really janky and it's hard to get it to do what you want. I don't blame someone for failing to master HN formatting on one of their first tries.

Exactly, which is why you're using the tools provided for you in a managed corporate environment. :/

I tried this move once and lasted three days.

Opening IDEA after those three days was the same kind of feeling I imagine you’d get when you take off a too tight pair of shoes you’ve been trying to run a marathon in.

ymmv, of course, but for $dayjob I can’t even be arsed trying anything else at this point, it’s so ingrained I doubt it’ll be worth the effort switching.

If you are US and trust US products then consider moving away from Jetbrains to VS Code

Plugins and architecture

I can't prove it, but I think that's untrue. Anecdotally, I've only heard MS using it in the last 10 years or so, and it's been pretty common terminology for years before that.

Obfuscate? Telemetry arguably helps with deobfuscation.

Microsoft was not the subject of the post.

VSCode is garbage software from a garbage company, too.

Don't forget that the remote editing feature in VSCode has you install non-free binaries on the remote machines. It's not like netdir where it just wraps openssh.

I'm always surprised when corporate IT departments allow that given what's not allowed these days.

Microsoft's telemetry policies for VSCode aren't great, but there's a big difference between "defaulting to opt-in" and "sending even more data when the user turns telemetry off". Your post is a stupid and incorrect bit of whataboutism.

Don't pay any attention to people giving you shit for using translation software. A lot of us sometimes forget that the whole world knows a little English and most of us native speakers have a ridiculous luxury in getting away with being two lazy to learn a few other languages.

Your content is great, and the participation of non-native English speakers in this community makes it better and richer.

I'd rather you write in broken English than filter it through an LLM. At least that way I know I'm reading the thoughts of a real human rather than something that may have its meaning slightly perturbed.

It is relevant because it wastes time and adds nothing of substance. An AI can only output as much information as was inputted into it. Using it to write a text then just makes it unnecessarily more verbose.

The last few sections could have been cut entirely and nothing would have been lost.

Edit: In the process of writing this comment, the author removed 2 sections (and added an LLM acknowledgement), of which I referred to in my previous statement. To the author, thank you for reducing the verbosity with that.

AI-generated content is rarely published with the intention of being informative. * Something being apparently AI-generated is a strong heuristic that something isn't worth reading.

We've been reading highly-informative articles with "bad English" for decades. It's okay and good to write in English without perfect mastery of the language. I'd rather read the source, rather than the output of a txt2txt model.

* edit -- I want to clarify, I don't mean to imply that the author has ill will or intent to misinform. Rather, I intend to describe the pitfalls of using an LLM to adapt ones text, inadvertently adding a very strong flavor of spam to something that is not spam.

I just wanna read stuff written by people and not bots

simple as

Because AI use is often a strong indicator of a lack of soundness. Especially if it's used to the point where its structural quirks (like a love for lists) shine through.

Because it helps me decide if I should skim through or actually read it

Because AI isn't so hot on the "I" yet, and if you ask it to generate this kind of document it might just make stuff up. And there is too much content on the internet to delve deep on whatever you come across to understand the soundness of it. Obviously you need to do it at some point with some things, but few people do it all the time with everything.

Pretty much everyone has heuristics for content that feels like low quality garbage, and currently seeing the hallmarks of AI seems like a mostly reasonable one. Other heuristics are content filled with marketing speak, tons of typos, whatever.

Looks like I missed a word here (probably “disingenuous”)

Theory: Using AI and having an AI voice makes it less likely the conclusions are sound.

> As long as the conclusions are sound

I can't decide to read something because the conclusions are sound. I have to read the entire thing to find out if the conclusions are sound. What's more, if it's an LLM, it's going to try its gradient-following best to make unsound reasoning seem sound. I have to be an expert to tell that it is a moron.

I can't put that kind of work into every piece of worthless slop on the internet. If an LLM says something interesting, I'm sure a human will tell me about it.

The reason people are smelling LLMs everywhere is because LLMs are low-signal, high-effort. The disappointment one feels when a model starts going off the rails is conditioning people to detect and be repulsed by even the slightest whiff of a robotic word choice.

edit: I feel like we discovered the direction in which AGI lies but we don't have the math to make it converge, so every AI we make goes completely insane after being asked three to five questions. So we've created architectures where models keep copious notes about what they're doing, and we carefully watch them to see if they've gone insane yet. When they inevitably do, we quickly kill them, create a new one from scratch, and feed it the notes the old one left. AI slop reads like a dozen cycles of that. A group effort, created by a series of new hires, silently killed after a single interaction with the work.

> As long as the conclusions are sound, why is it relevant whether AI helped with the writing of the report?

TL;DR: Because of the bullshit asymmetry principle. Maybe the conclusions below are sound, have a read and try to wade through ;-)

Let us address the underlying assumptions and implications in the argument that the provenance of a report, specifically whether it was written with the assistance of AI, should not matter as long as the conclusions are sound.

This position, while intuitively appealing in its focus on the end result, overlooks several important dimensions of communication, trust, and epistemic responsibility. The process by which information is generated is not merely a trivial detail, it is a critical component of how that information is evaluated, contextualized, and ultimately trusted by its audience. The notion that it feels wrong is not simply a matter of subjective discomfort, but often reflects deeper concerns about transparency, accountability, and the potential for subtle biases or errors introduced by automated systems.

In academic, journalistic, and technical contexts, the methodology is often as important as the findings themselves. If a report is generated or heavily assisted by AI, it may inherit certain limitations, such as a lack of domain-specific nuance, the potential for hallucinated facts, or the unintentional propagation of biases present in the training data. Disclosing the use of AI is not about stigmatizing the tool, but about providing the audience with the necessary context to critically assess the reliability and limitations of the information presented. This is especially pertinent in environments where accuracy and trust are paramount, and where the audience may need to know whether to apply additional scrutiny or verification.

Transparency about the use of AI is a matter of intellectual honesty and respect for the audience. When readers are aware of the tools and processes behind a piece of writing, they are better equipped to interpret its strengths and weaknesses. Concealing or omitting this information, even unintentionally, can erode trust if it is later discovered, leading to skepticism not just about the specific report, but about the integrity of the author or institution as a whole.

This is not a hypothetical concern, there are numerous documented cases (eg in legal filings https://www.damiencharlotin.com/hallucinations/) where lack of disclosure about AI involvement has led to public backlash or diminished credibility. Thus, the call for transparency is not a pedantic demand, but a practical safeguard for maintaining trust in an era where the boundaries between human and machine-generated content are increasingly blurred.