An itch that I have been trying to scratch for a while: I need a scripting language that I can integrate into a web app, and I can debug from the web app. It doesn't need to be complex or with curly braces, actually the simple the better.
Three days with AI and FluidScript is born. The language is minimalistic, the keywords are simple, and you can stop execution at any line and you get the whole context back, later you can continue from that point passing the context, and it continues as if it hasn't stopped. Full host integration and a Monaco editor example that shows how to do break points.
Dreamed it for years, now, once I have a clear concept and design, AI type it on 3 days.
Yup!. Over 30 years experience here. Landed on the same place with the added feeling of "We were always intended to develop the judgement of what makes sense and maintainable, where is the surprise here?"
People want the "Fluent Text Generator" to think for them because its marketing calls it AI. At the end is just a tensor collection, probabilistically backtracked adjusted, text generator. It does produces fluency in the answer, but the fact that it sounds correct, to the point of passing a compiler approval, has nothing to do with been well thought out.
I think it's easier to make your point without the baggage of the mechanism assertions.
The code is just wrong a lot. By denying the possibility of emergent phenomemona or model interiority you just hand the hypsters an easy point to score.
The code is just wrong, all the time. You're right about the part that matters and that we can measure.
Except... it does produce solid code in many scenarios. A lot of the time we are just applying the same patterns and tools to a specific problem. If you understand the problem, agents can and are producing quality code.
I do a lot with orchestrating cloud infra, provisioning and deprovisioning workflows, lifecycle management, updating CMDBs, access management, deployment automation, auditing configurations, and some APIs in front. The code is boring, mostly glue. The agents nail it quite consistently, especially with some guidance.
I do know what I want out of it, and that is deterministic quality code to solve specific business problems, so that we don't have people (or their agents) running amok.
it does do good code sometimes. it does both. we're trying to figure out how to net that out to good overall. right now it's a bit tricky because we didn't build our world for the code to be really good and then disasterously bad all the sudden and then good for a while again and it seems like someone is programming a slot machine.
just to test the tech. No real usage other than for the fun of it.
I did port stb_image from C to Jai which i was able to fully verify and harden and that one ill give more use. Im also using the same workflow system to perform agentic translation of a game i work with from english to various other languages, the results are far better than the commercial "human" translation services we tested. And i also use it to fix OCR issues on PDF books im ocr-ing for a data pipeline. This kind of workflow/wide agent swarm system is rather useful for many things where you want to "apply" the same prompts across a whole codebase or just in parallel.
i decided to do an experiment and try to run an LLM in my old 2013 MBP. i7, 16 gb mem, 1 tb hd.
Installed Linux mint Xfce Edition for lightness, installed ollama, start to test different models. Gemma4 e4b runs perfectly fine, exposed it to the network, connected to it with my current notebook and use vs code codex to start to run inference.
For about 30 minutes of bliss, this setup work at a reasonable speed... then the MBP shut it self down. It was so hot that it trigger the safety mechanism, the fans sounded like the laptop was about to take off.
I though on leaving it on inside the fridge, but then the WIFI wouldn't reach.
On the other hand, my wife saw all this and offer to buy me an M5... the experiment didn't work as intended, but it did work.
On a laptop that old it might be worth opening it up to blow all the dust out with a compressor or air duster. I’ve often found that to work wonders on old MacBook Pros.
The other issue is that unless the battery has been replaced relatively recently its charging efficiency may not be that great and the high load being placed on it might be causing it to get hotter than it would have done when new.
I was all for trying it until I saw this in the License Agreement:
"4.1. Zed's Use of Customer Data
Customer hereby grants Zed a non-exclusive, worldwide, royalty-free, fully paid-up, non-sublicensable (except to service providers and Customer’s designees), non-transferable (except as set forth in Section 15.1) right to use, copy, store, disclose, transmit, transfer, display, modify, create derivative works from, collect, access, store, host, or otherwise process (“Process”) any materials that Customer inputs into or otherwise makes available to the Service (including prompts and other written content) (collectively, “Customer Data”) solely: (a) to perform its obligations set forth in the Terms, including its Support obligations as applicable; (b) to derive and generate Telemetry (see Section 4.4); and (c) as necessary to comply with applicable Laws. Except as required by applicable Laws, Zed will not provide Customer Data to any person or entity other than Customer’s designees (including pursuant to Section 7) or service providers."
Sorry, no I don't agree to make my source code and the product I am working to give you "non-exclusive, worldwide, royalty-free, fully paid-up, non-sublicensable (except to service providers and Customer’s designees), non-transferable (except as set forth in Section 15.1) right to use, copy, store, disclose, transmit, transfer, display, modify, create derivative works from, collect, access, store, host, or otherwise process (“Process”) any materials that Customer inputs into or otherwise makes available to the Service (including prompts and other written content)"
Aren't you forgetting the part that says "solely: (a) to perform its obligations set forth in the Terms, including its Support obligations as applicable; (b) to derive and generate Telemetry (see Section 4.4); and (c) as necessary to comply with applicable Laws. Except as required by applicable Laws, Zed will not provide Customer Data to any person or entity other than Customer’s designees (including pursuant to Section 7) or service providers."
Seems like legalese to be able to take that data for support reasons, telemetry, and local laws that require that data be sent to them. I think ignoring this portion is a little uncharitable to them.
>Aren't you forgetting the part that says "solely: (a) to perform its obligations set forth in the Terms, including its Support obligations as applicable; (b) to derive and generate Telemetry (see Section 4.4); and (c) as necessary to comply with applicable Laws
None of the above I like, and (a) is so vague as to be useless, including if you read the obligations.
>Except as required by applicable Laws, Zed will not provide Customer Data to any person or entity other than Customer’s designees (including pursuant to Section 7) or service providers."
Companies still do it all the time despite "applicable laws". And when the company is sold, all bets are off.
I'd rather they don't get, or keep, any to begin with.
The telemetry section of the TOS explicitly clarifies that that does not restrict their ability to use the data that is sent to them.
> Customer may configure the Software to opt out of the collection of certain Telemetry Processed locally by the Software itself, but Zed may still collect, generate, and Process Telemetry on Zed’s servers.
Note that they have (or did have, I haven't used their editor in awhile) an AI tab completion feature... it's safe to assume that all of the code you edit is sent to them at least when that is enabled.
>The telemetry section of the TOS explicitly clarifies that that does not restrict their ability to use the data that is sent to them.
Hopefully it does restrict them being sent to them in the first place.
I also found there are a couple of "Chromium" style builds.
>Note that they have (or did have, I haven't used their editor in awhile) an AI tab completion feature... it's safe to assume that all of the code you edit is sent to them at least when that is enabled.
There's also an option to turn ai features off. At which point of course, nvim is just as good :)
What I understand reading this is that if you use their online services, incl AI-agents, llm based tab-completion, auto-updates etc, you send data to their servers, and on that part they run analytics. Frankly, this is what I would expect anyway, ie if I disable telemetry locally, it would affect what I do locally, ie no data about how I use my software locally would leave the machine, but if I sent data to some server I would not expect people not to run analytics on their servers.
> AI tab completion feature... it's safe to assume that all of the code you edit is sent to them
Yes, this is quite obvious, how else could they provide AI tab completion? I hope anybody understands this before using sth like this. They do specify that "[...] telemetry expressly does not include Customer Data" though.
> They do specify that "[...] telemetry expressly does not include Customer Data" though.
Yes and no. They first grant themselves a license "to derive and generate Telemetry" from the users copyrighted material, something that they only need if they're deriving it from the actual creative works the customer updates, and not just the metadata about them.
And they define telemetry extremely broadly, effectively "anything useful for lawful business purposes except customer data".
So this agreement would seem to cover things like "an update to an AI model trained off of your code" or even "an AI summary of what you're working on and any relevant business information contained therein". As long as they process it to something new, it's not "customer data" (a term defined narrowly in the agreement). I don't expect that they are doing that, but I think they've given themselves permission to. The agreement is far too broad.
I agree that I expect that they are deriving metadata, and would expect that regardless of this agreement, but this agreement doesn't seem necessary for that.
I was willing to give it another go. Now I read on this thread that it installs tons of node packages (so much for Rust native code) and even Go packages, and gets many extra processes running along with it.
TL;DR: Mix of language tooling, unsigned proprietary blobs, corrupted and/or GLIBC-dependent files, redundant copies of already-installed executables. The Node packages especially are able to run scripts on install. Personal preference aside, might also create issues with security laws, certifications. All without user consent.
Issues opened in January and June 2024. They've been rejected, closed, and opened a couple times since then. No changes directly improving this yet as of April 2026.
Personally, I think even if they eventually fix this, given the attitude shown towards their users' machines, I should probably just use an editor where I don't have to worry about it.
This is a bad faith take. The terms are modifiable without the customer's consent or knowledge so "pursuant to these terms" is meaningless.
No one needs all those rights to do what this block says it's going to do. Any one would require that block to be changed in any contract between equals. But this is a contract of adhesion, so it's uncharitable for you to demand charity where they withhold their charity
Can you cite the passage that authorizes Zed to modify the terms without the user's consent? Before I retired, my job was, inter alia, writing software licenses. I was GC for a tech company. I'd like to validate what you're saying, bc I'm the author of a Zed plugin and I wrote a language grammar that another plugin uses.
I don't use Zed, but I occasionally consider switching.
That's not what 16.4 says, although I can understand how you'd think that (and personally, I would not have written 16.4 this way, and I've had a longstanding issue with a lot of software license phrasing; used to find all sorts of loopholes in partner agreements even with enormous companies like AT&T). 16.4 warns the user that terms might change, those terms will be available, and continued use of the product after the new terms are promulgated will constitute acceptance of the terms.
When new terms come out, you're forcibly signed out of the Zed service and have to agree to new ones before using the service again.
It's most likely the case that this clause was written so that someone using the service, when offered new terms, can't say "nah, Imma stick with the old terms" and then claim Zed aren't allowed to revoke the license.
I had the same thought but if you chase down the definition of "Telemetry" as well as unilateral amendment rights pointed out in sibling comments, there's some broad authority implied.
Probably 'we reserve the right to train our next version of smart autocomplete based on the text you send to the current version of smart autocomplete'
Which is not different in kind from “we use your source code to improve our products” and is functionally identical to “we own your output because you use our editor.”
How do people continually fall for this. Refusing to look at the playbook that has been run time and time again and then getting offended when it is too late.
They don't. Paragraph 4.2, "Customer's Ownership of Output" [0]. I recite verbatim below for the sake of clarity.
These are about processing the data, not owning it. They need to process the data eg to provide llm-based tab-completion. A completion is derivative work, and it is also owned by the customer, as it says below.
> The Service may generate specifically for, and make available to, Customer text and written content based on or in response to Customer Data input into the Service (collectively, “Output”), including through the use of technologies that incorporate or rely upon artificial intelligence, machine learning techniques, and other similar technology and features. As between the Parties, to the greatest extent permitted by applicable Laws, Customer owns all Output and Zed hereby irrevocably assigns to Customer all right, title, and interest in and to the Output that Zed may possess. For the avoidance of doubt, Zed and its AI Providers will not retain or use Customer Data for the purpose of improving or training the Service or any AI Provider products, except to the extent Customer explicitly opts-in on Zed’s specific feature to allow training and/or such improvement (such as fine-tuning) and is solely for the benefit of Customer.
I had a long conversation about this with Gemini this morning. I described the telemetry practices and Gemini told me about all the local and federal laws that were being violated. Then I mentioned it was telemetry, and it turned on a dime and said it was fine because the user agreed to it in the TOS. Disgusting.
I'm not a lawyer, but the only part of this that seems objectionable is the "telemetry" bit, the rest of it basically seems to say "we can use things you send us to do the things you asked us to do, including for support purposes. We can comply with the law as necessary (e.g. responding to warrants)".
That said the definition of "telemetry" is so broad that I think it would include training a LLM and the like. Telemetry is defined in section 4.4 as
> Zed may collect, generate, and Process information, including technical logs, metrics, and data and learnings, related to the Software and Service (“Telemetry”) to improve and support the Services and for other lawful business purposes.
I guess that it's so opaque is also objectionable. Contracts don't have to be inscrutable.
I understand that differently. The last part of the statement is important I think. It reads as if this grants them the right to "Process" "Customer Data":
1. to perform its obligations, including support obligations,
2. to perform telemetry,
3. when required by law.
Can't you just build the source and run it without giving them any information? Or does the editor itself require information and phone home while running?
That license agreement covers their online service if you opt in to use it. This reads like every ToS for a SaaS product I’ve read in ages. If you don’t want to use their paid service, don’t create an account and configure it in Zed, and then this doesn’t apply to you.
I’m not associated with Zed in any way except as a user. That said, this is a bonkers misinterpretation of their ToS. You don’t grant them the right to eat all your data just by using the editor. You do give them the right to process your data if you sign up for an account and opt in to send them all your data for processing.
> The Terms of Service and the open source editor are separate. Zed's editor is licensed under GPL v3 (with Apache 2.0 for certain components). Those licenses govern the software itself and haven't changed. The Terms of Service apply when you create a Zed account and use our hosted services. If you use the editor without signing in, the open source licenses are the only thing that applies, and the new terms codify this more clearly than the old.
so something that 95% of the users of Zed will end up doing? You're arguing this is a good thing because there is a "choice" in how the company gets to fuck you over?
> so something that 95% of the users of Zed will end up doing?
Will they? I downloaded it for a test run, and there was no pressure to create a Zed account. I got the impression that it's something you'd do if you wanted to use their cloud AI services, and I can't really see why you'd want a third party involved instead of just bringing your own subscription to your favourite model.
Not sure what you assume I said or argued. Grandparent wrote a question about how can an open source project have ToS (which is a totally legitimate question if they think that the ToS are for the open source project) and I copy pasted the response of somebody from zed team that says that the ToS are about the subscription services only.
The data related part of the ToS for the subscription are the most basic thing you would expect in a standard data processing agreement type of ToS where you have somebody else process your data on your behalf. I don't see any "fuck you over" in them.
This language fits common SaaS templates. Let me illustrate by removing chunks and labeling them:
Customer hereby grants Zed
{{ broad list of rights }}
solely:
{{ for these purposes }}
IANAL, but the term "solely" seems essential to understanding this. Pivotal. As in "if you get it wrong, you'll be wearing a tinfoil hat" essential.
Also see "4.4. Telemetry: .... For avoidance of doubt, Telemetry expressly does not include Customer Data."
My two cents: I'm not worried about Zed's contract here. Much more important to pay attention to when your data goes to third-party AI providers: read _their_ contract language.
Meta-comment: Don't let a well-meaning comment like the above trigger a panic. Better to get familiar with typical contracts and/or build your personal network for legal advice.
P.S. Look out for shameless legal slop on the Web, I "promise.legal" it is out there.
Yeah this really makes you wonder... now, they are to best of my knowledge fantastic people, so I guess some lawyer slipped this in. I would love to hear some clarification and will always give them benefit of the doubt.
So far they've been great and product is fantastic as you know.
> now, they are to best of my knowledge fantastic people,
They could be the next "don't do evil" people but practice shows that doesn't last for long. And then the messy license terms become very handy for what comes next.
If they went to the trouble to specify all of their rights over your data, their glossing over about what they can do with it is a solid reason to push for complete clarity or pull out completely.
wouldn't touch this editor with a 10 foot pole. no i am not giving them the rights to use my code for their product. hell no. thanks for highlighting this. people need to making a bigger deal about it. maybe they will rethink their tos.
I encourage you to mail them specifically about this concern.
If it is a matter of communication, they can fix and clarify it. If it's genuinely scummy, they can change their approach now that they know they were caught in the act
The comment above has started a sh-tstorm. Please, slow down and learn about the details before jumping to conclusions. Most of us here did NOT "go pro" in the law. [1] For those that want to educate themselves, you could do worse than immediately leave HN and go _learn_:
If anyone has good detailed resources that are free, please add.
[1] IANAL but I wasn't that far from going down that path. I've worked for a legal-tech startup, did really well in an undergrad Constitutional Law class, incorporated several small companies, managed lots of contractor agreements. So, I know from experience: legal language is weird and specific in ways you may not realize. So be intellectually humble and defer judgment until you talk to a legal expert. Hopefully people more experienced than I can weigh in with more specifics.
If you need an education in law to be able to trust a business isn't trying to steal from you, then maybe you just shouldn't trust that business at all.
Especially for something like a code editor, where plenty of less-shady competitors are available.
1. Mandatory arbitration by default. You waive your right to a jury trial and cannot join class action lawsuits. You have only 30 days to opt out after agreeing to the terms.
2. 1-year statute of limitations. You must file any legal claim within one year, which is much shorter than the default period under most state laws.
3. Zed can terminate your account at any time, for any reason, with no liability. No notice is required, and they owe you nothing if they pull the plug.
4. Autocomplete sends your code to AI providers in the background unless you turn it off. Worth knowing if you're working on proprietary or sensitive code.
5. No guarantees about data retention. If your payment lapses, they reserve the right to delete your account and all associated data with no liability.
6. All fees are non-refundable except where required by law, with one narrow exception for disagreeing with modified terms.
7. Zed can modify the terms at any time. For existing users, material changes take effect after just 30 days, and continued use counts as acceptance.
8. Zed can use your name, logo, and brand in their marketing without asking. You'd have to send a written request to stop them.
9. No warranties whatsoever. The service is "as is", they disclaim all responsibility for errors, data loss, or AI-generated output being inaccurate or harmful.
10. Liability is capped very low, at most, whatever you paid in the last 12 months, or $100, whichever is higher.
What I do is to have two things, a simple editor, I use helix for normal editing. And in a second terminal a docker container solution where I put opencode or claude in https://git.jeena.net/jeena/agent-container
By sandbox you mean limit to certain files, certain actions, or both?
I've been wanting to look into better emacs integration for agents. Imagine an agent making direct elisp function calls, or using macros... One could limit which functions are allowed to run similar to how cli harnesses work, but plug straight into LSP and etc.
* The developer stop been a developer, and becomes a product designer with high technical skills.
* This is a different set of skills than than a developer or a product owner currently have. It is a mix of both, and the expectations of how agentic development works need to be adjusted.
* Agents will behave like junior developers, they can type very fast, and produce something that has a high probability to work. They priority will be to make it work, not maintainability, scalability, etc. Agents can achieve that if you detail how to produce it.
* The working with an agent feels more like mentoring the AI than ask and receive.
* When I start to work on a product that will be vibe coded, I need to have clear in my head all the user stories, code architecture, the whole system, then I can start to tell the agent what to build, and correct and annotate in the md files the code quality decisions so it remembers them.
* Use TDD, ask the agent to create the tests, and then code to the test. Don't correct the bugs, make the agent correct them and explain why that is a bug, specially with code design decisions. Store those in AGENTS.md file at the root of the project.
There are more things that can be done to guide the agent, but I need to have clear in an articulable way the direction of the coding. On the other side, I don't worry about implementation details like how to use libraries and APIs that I am not familiar with, the agent just writes and I test.
Currently I am working on a product and I can tell you, working no more than 10 hours a week (2 hours here, 3 there, leave the agent working while I am having dinner with family) I am progressing at I would say 5 to 10 times faster than without it. So, yeah it works, but I had to adjust how I do my job.
This forum (HN) attracts certain population that wants to do things, to understand, to share relatively well based opinions and have a discussion.
But look around, look at the new hires in the other departments. And by new I mean young, in their 20. A lot of them welcome this kind of things, they evaluate by popularity and likes. The marketing begin the AI bubble knows this, and so it pushes for it. Make it popular is more important than make it useful, because there is a tipping point were is popular enough that we capitulate.
Did they, though? Polling fairly consistently shows that people don’t _like_ this stuff, and there’s some evidence that the more familiar with it they become the less they like it. I think Microsoft et al were betting on people liking it (that was certainly their thinking with Clippy, too) but that doesn’t seem to be working out for them.
My sister is ripping her hair out dealing with the interns at he job being extremely tech illiterate with anything thats not app-ified. Many don't know what files are, and are needing run through computer basics because everything they have used has anything technical abstracted away. The post-iPhone generation just wants there hand held and anything technical scares them. Microsoft Bob was just too ahead of the curve.
Not only have the people changed but it is the belief of the elites at the top that humanity is entering a new era of hack-ability. They want to use these AI systems to rewrite humanity into their vision of the future.
Yuval Noah Harari talking about how the new "gods" are the data-centers and how free will is dead in the age of AI.
https://youtu.be/QuL3wlodJC8
Well, for the purpose of this conversation the people at the top of the food chain believe free will exists. They also believe that they can eliminate it with AI and Biomedical manipulation.
The while message of the article is to trash talk the departing editor accusing her of political left bias... which in it self (the trash talking) is a political statement from the conservative side.
To the author of the article: you are no better than her...
Can you reconcile that view with the paragraph at the end?
>That doesn't mean the editor needs to be apolitical or that there's no role for SciAm to chime in on social justice issues in an informed manner, with the requisite level of humility and caution. It simply means that Scientific American needs to get back to its roots—explaining the universe's wonders to its readers, not lecturing them about how society should be ordered or distorting politically inconvenient findings.
No need to reconcile because one thing does not excuse the other.
If I go complaining that you go around beating people up, and that is why I will go and beat you up, and at the end I claim that it is ok because I agree with hitting people is ok doesn't excuse my action.
Also, stating the obvious (SA needs to get back to its roots) serves in this case as a straw man argument, the point was how bad an inexcusable was the editor behavior, not what the roots of SA should be.
This article is closer to the son of the president in the "Don't look up" movie than anything else. It tries to push the previous editor to a square of just do scientific work... but there is a point, in defense of the editor, where people claiming that the earth is flat need to be push back. Objective truth needs to prevail regardless of how people feel about it politically, and it is ok, in my book, to defend that
The point of the article is that SA can't sacrifice science to push propaganda. That's it.
Like this point of yours:
>but there is a point, in defense of the editor, where people claiming that the earth is flat need to be push back. Objective truth needs to prevail regardless of how people feel about it politically, and it is ok, in my book, to defend that
Is true, for the article, not for the editor. That's his whole point.
1. Defending science as objective truth is not propaganda, so the editor did not engage un such.
2. The article it self is not about science, but it is weasel propaganda on it self because accusing of propaganda to the editor is a form of propaganda that is presented as "reasonable" but the intended effect is to try to call propaganda what is not.
I genuinely have no clue what you are talking about.
>1. Defending science as objective truth is not propaganda, so the editor did not engage un such.
A headline like "Why the Term 'JEDI' Is Problematic for Describing Programs That Promote Justice, Equity, Diversity and Inclusion." has absolutely, completely nothing to do with "defending science as objective truth". Within the article political views are pushed, with science being largely irrelevant to the case.
>2. The article it self is not about science, but it is weasel propaganda on it self because accusing of propaganda to the editor is a form of propaganda that is presented as "reasonable" but the intended effect is to try to call propaganda what is not.
The article does not attempt to be seen "about science" at any point. It's not weaseling at any point, it's point is made very clearly. It even makes the point that propaganda isn't necessarily wrong
Three days with AI and FluidScript is born. The language is minimalistic, the keywords are simple, and you can stop execution at any line and you get the whole context back, later you can continue from that point passing the context, and it continues as if it hasn't stopped. Full host integration and a Monaco editor example that shows how to do break points.
Dreamed it for years, now, once I have a clear concept and design, AI type it on 3 days.
reply