Atlassian contains the data of over 300k organizations. Companies of all sizes use their products, including free users, small teams, large organizations, and enterprises.
Starting Aug 17th, if your company has not opted out of \u201cData Contribution\u201d, Atlassian will use your company\u2019s data to train their AI products (\u201cRovo\u201d).
The intrinsic value of the data residing in Atlassian\u2019s products is uniquely high. Additionally, how Atlassian is rolling out Data Contribution is hard to view favorably.
> On intrinsic value:
Atlassian has several product offerings but their main two are Jira and Confluence. Confluence is a documentation platform containing the knowledge base of many companies. Jira, a ticketing/product system, contains a temporally organized record of a company's operational processes and their execution steps for delivering their products. Many Jira instances contain long term execution intentions towards an overarching company strategy.
The synergy of both of those, the knowledge base and tasks/intentions, is impressively valuable. For many organizations, the completeness of this data in both of these tools is high. Additionally, the recency and freshness of the data is near real time. The pairing of both Jira and Confluence data adds incredible contextual relevance to understanding the company.
Continuing, the very position and nature of these tools, be it their ease of integrations, the fluidity of adding attachments, the social aspect of the platforms, the requisite requirement of using the tools in many development processes, etc. has allowed these platforms to accumulate a large amount of intellectual property from companies.
> On the rollout:
There are two types of data to be collected and trained on, 1) \u201cMetadata\u201d and 2) \u201cData\u201d. The only way a company can opt-out of both is if they are on an Enterprise subscription, otherwise Data opt-out is a manual slider and Metadata is always contributed. The problem with Atlassian Enterprise is its inaccessibility. Some SaaS services (GitHub, for ex) - allow smaller organizations to easily self-sign for Enterprise. It is more costly per seat but organizations can get access to the same features as enterprises. Atlassian does not have this level of accessibility, a company has to contact sales to discuss an Enterprise account. Even then, the cutoffs for user counts are significantly higher (800+ users is my understanding, but there are probably more accurate numbers).
Atlassian has made an effort to separate the types of data into Metadata and Data - but their definition of Metadata is not metadata in the classical definition. Their \u201cMetadata\u201d includes 1) numeric fields like story points, dates (which they call numbers in their docs), SLAs, etc. 2) computed features on your data (similarity scores, readability scores, etc.), and more. Those are stored as \u201cMetadata\u201d for use.
> Extrapolating:
A weird corporate welfare forms. We essentially have partially-opt-outable organizations \u201ccontributing\u201d their organizational processes + IP in some anonymized form to Atlassian for Rovo development, while the largest and most successful enterprises are not having to share their same value back. Many small organizations make a market for themselves by being first to market, filling a niche, and building responsive products faster than larger firms.
While Atlassian will anonymize and remove PII and specifics, where on the sliding scale of reproducible business strategy process will we land \u2013 New York Times + ChatGPT regurgitation? Which organizations will be able to capitalize the most on that trained information coming from thousands of smaller organizations?
> TLDR - Atlassian training on your data via \u201cData Contribution\u201d creates a privacy and IP concern, and their policy rollout results in small organizations contributing their knowledge and process to large organizations without commensurate contribution in return."},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["atlassian"],"value":"Atlassian \"Data Contribution\" \u2013 Privacy and Welfare"}},"_tags":["story","author_yells_jovially","story_48540151","ask_hn"],"author":"yells_jovially","children":[48540202],"created_at":"2026-06-15T12:11:56Z","created_at_i":1781525516,"num_comments":1,"objectID":"48540151","points":4,"story_id":48540151,"story_text":"> For the uninitiated:
Atlassian contains the data of over 300k organizations. Companies of all sizes use their products, including free users, small teams, large organizations, and enterprises.
Starting Aug 17th, if your company has not opted out of \u201cData Contribution\u201d, Atlassian will use your company\u2019s data to train their AI products (\u201cRovo\u201d).
The intrinsic value of the data residing in Atlassian\u2019s products is uniquely high. Additionally, how Atlassian is rolling out Data Contribution is hard to view favorably.
> On intrinsic value:
Atlassian has several product offerings but their main two are Jira and Confluence. Confluence is a documentation platform containing the knowledge base of many companies. Jira, a ticketing/product system, contains a temporally organized record of a company's operational processes and their execution steps for delivering their products. Many Jira instances contain long term execution intentions towards an overarching company strategy.
The synergy of both of those, the knowledge base and tasks/intentions, is impressively valuable. For many organizations, the completeness of this data in both of these tools is high. Additionally, the recency and freshness of the data is near real time. The pairing of both Jira and Confluence data adds incredible contextual relevance to understanding the company.
Continuing, the very position and nature of these tools, be it their ease of integrations, the fluidity of adding attachments, the social aspect of the platforms, the requisite requirement of using the tools in many development processes, etc. has allowed these platforms to accumulate a large amount of intellectual property from companies.
> On the rollout:
There are two types of data to be collected and trained on, 1) \u201cMetadata\u201d and 2) \u201cData\u201d. The only way a company can opt-out of both is if they are on an Enterprise subscription, otherwise Data opt-out is a manual slider and Metadata is always contributed. The problem with Atlassian Enterprise is its inaccessibility. Some SaaS services (GitHub, for ex) - allow smaller organizations to easily self-sign for Enterprise. It is more costly per seat but organizations can get access to the same features as enterprises. Atlassian does not have this level of accessibility, a company has to contact sales to discuss an Enterprise account. Even then, the cutoffs for user counts are significantly higher (800+ users is my understanding, but there are probably more accurate numbers).
Atlassian has made an effort to separate the types of data into Metadata and Data - but their definition of Metadata is not metadata in the classical definition. Their \u201cMetadata\u201d includes 1) numeric fields like story points, dates (which they call numbers in their docs), SLAs, etc. 2) computed features on your data (similarity scores, readability scores, etc.), and more. Those are stored as \u201cMetadata\u201d for use.
> Extrapolating:
A weird corporate welfare forms. We essentially have partially-opt-outable organizations \u201ccontributing\u201d their organizational processes + IP in some anonymized form to Atlassian for Rovo development, while the largest and most successful enterprises are not having to share their same value back. Many small organizations make a market for themselves by being first to market, filling a niche, and building responsive products faster than larger firms.
While Atlassian will anonymize and remove PII and specifics, where on the sliding scale of reproducible business strategy process will we land \u2013 New York Times + ChatGPT regurgitation? Which organizations will be able to capitalize the most on that trained information coming from thousands of smaller organizations?
> TLDR - Atlassian training on your data via \u201cData Contribution\u201d creates a privacy and IP concern, and their policy rollout results in small organizations contributing their knowledge and process to large organizations without commensurate contribution in return.","title":"Atlassian \"Data Contribution\" \u2013 Privacy and Welfare","updated_at":"2026-06-15T18:32:00Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"yells_jovially"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"> For the uninitiated:\nAtlassian contains the data of over 300k organizations. Companies of all sizes use their products, including free users, small teams, large organizations, and enterprises.\nStarting Aug 17th, if your company has not opted out of \u201cData Contribution\u201d, Atlassian will use your company\u2019s data to train their AI products (\u201cRovo\u201d).\nThe intrinsic value of the data residing in Atlassian\u2019s products is uniquely high. Additionally, how Atlassian is rolling out Data Contribution is hard to view favorably.\n> On intrinsic value:\nAtlassian has several product offerings but their main two are Jira and Confluence. Confluence is a documentation platform containing the knowledge base of many companies. Jira, a ticketing/product system, contains a temporally organized record of a company's operational processes and their execution steps for delivering their products. Many Jira instances contain long term execution intentions towards an overarching company strategy.\nThe synergy of both of those, the knowledge base and tasks/intentions, is impressively valuable. For many organizations, the completeness of this data in both of these tools is high. Additionally, the recency and freshness of the data is near real time. The pairing of both Jira and Confluence data adds incredible contextual relevance to understanding the company. \nContinuing, the very position and nature of these tools, be it their ease of integrations, the fluidity of adding attachments, the social aspect of the platforms, the requisite requirement of using the tools in many development processes, etc. has allowed these platforms to accumulate a large amount of intellectual property from companies.\n> On the rollout:\nThere are two types of data to be collected and trained on, 1) \u201cMetadata\u201d and 2) \u201cData\u201d. The only way a company can opt-out of both is if they are on an Enterprise subscription, otherwise Data opt-out is a manual slider and Metadata is always contributed. The problem with Atlassian Enterprise is its inaccessibility. Some SaaS services (GitHub, for ex) - allow smaller organizations to easily self-sign for Enterprise. It is more costly per seat but organizations can get access to the same features as enterprises. Atlassian does not have this level of accessibility, a company has to contact sales to discuss an Enterprise account. Even then, the cutoffs for user counts are significantly higher (800+ users is my understanding, but there are probably more accurate numbers).\nAtlassian has made an effort to separate the types of data into Metadata and Data - but their definition of Metadata is not metadata in the classical definition. Their \u201cMetadata\u201d includes 1) numeric fields like story points, dates (which they call numbers in their docs), SLAs, etc. 2) computed features on your data (similarity scores, readability scores, etc.), and more. Those are stored as \u201cMetadata\u201d for use. \n> Extrapolating:\nA weird corporate welfare forms. We essentially have partially-opt-outable organizations \u201ccontributing\u201d their organizational processes + IP in some anonymized form to Atlassian for Rovo development, while the largest and most successful enterprises are not having to share their same value back. Many small organizations make a market for themselves by being first to market, filling a niche, and building responsive products faster than larger firms. \nWhile Atlassian will anonymize and remove PII and specifics, where on the sliding scale of reproducible business strategy process will we land \u2013 New York Times + ChatGPT regurgitation? Which organizations will be able to capitalize the most on that trained information coming from thousands of smaller organizations?\n> TLDR - Atlassian training on your data via \u201cData Contribution\u201d creates a privacy and IP concern, and their policy rollout results in small organizations contributing their knowledge and process to large organizations without commensurate contribution in return."},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["atlassian"],"value":"Atlassian \"Data Contribution\""}},"_tags":["story","author_yells_jovially","story_48522482","ask_hn"],"author":"yells_jovially","created_at":"2026-06-13T23:21:20Z","created_at_i":1781392880,"num_comments":0,"objectID":"48522482","points":2,"story_id":48522482,"story_text":"> For the uninitiated:\nAtlassian contains the data of over 300k organizations. Companies of all sizes use their products, including free users, small teams, large organizations, and enterprises.\nStarting Aug 17th, if your company has not opted out of \u201cData Contribution\u201d, Atlassian will use your company\u2019s data to train their AI products (\u201cRovo\u201d).\nThe intrinsic value of the data residing in Atlassian\u2019s products is uniquely high. Additionally, how Atlassian is rolling out Data Contribution is hard to view favorably.\n> On intrinsic value:\nAtlassian has several product offerings but their main two are Jira and Confluence. Confluence is a documentation platform containing the knowledge base of many companies. Jira, a ticketing/product system, contains a temporally organized record of a company's operational processes and their execution steps for delivering their products. Many Jira instances contain long term execution intentions towards an overarching company strategy.\nThe synergy of both of those, the knowledge base and tasks/intentions, is impressively valuable. For many organizations, the completeness of this data in both of these tools is high. Additionally, the recency and freshness of the data is near real time. The pairing of both Jira and Confluence data adds incredible contextual relevance to understanding the company. \nContinuing, the very position and nature of these tools, be it their ease of integrations, the fluidity of adding attachments, the social aspect of the platforms, the requisite requirement of using the tools in many development processes, etc. has allowed these platforms to accumulate a large amount of intellectual property from companies.\n> On the rollout:\nThere are two types of data to be collected and trained on, 1) \u201cMetadata\u201d and 2) \u201cData\u201d. The only way a company can opt-out of both is if they are on an Enterprise subscription, otherwise Data opt-out is a manual slider and Metadata is always contributed. The problem with Atlassian Enterprise is its inaccessibility. Some SaaS services (GitHub, for ex) - allow smaller organizations to easily self-sign for Enterprise. It is more costly per seat but organizations can get access to the same features as enterprises. Atlassian does not have this level of accessibility, a company has to contact sales to discuss an Enterprise account. Even then, the cutoffs for user counts are significantly higher (800+ users is my understanding, but there are probably more accurate numbers).\nAtlassian has made an effort to separate the types of data into Metadata and Data - but their definition of Metadata is not metadata in the classical definition. Their \u201cMetadata\u201d includes 1) numeric fields like story points, dates (which they call numbers in their docs), SLAs, etc. 2) computed features on your data (similarity scores, readability scores, etc.), and more. Those are stored as \u201cMetadata\u201d for use. \n> Extrapolating:\nA weird corporate welfare forms. We essentially have partially-opt-outable organizations \u201ccontributing\u201d their organizational processes + IP in some anonymized form to Atlassian for Rovo development, while the largest and most successful enterprises are not having to share their same value back. Many small organizations make a market for themselves by being first to market, filling a niche, and building responsive products faster than larger firms. \nWhile Atlassian will anonymize and remove PII and specifics, where on the sliding scale of reproducible business strategy process will we land \u2013 New York Times + ChatGPT regurgitation? Which organizations will be able to capitalize the most on that trained information coming from thousands of smaller organizations?\n> TLDR - Atlassian training on your data via \u201cData Contribution\u201d creates a privacy and IP concern, and their policy rollout results in small organizations contributing their knowledge and process to large organizations without commensurate contribution in return.","title":"Atlassian \"Data Contribution\"","updated_at":"2026-06-13T23:32:38Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"mononclemich"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"An isolation failure in an Atlassian Rovo-orchestrated environment allowed sessions belonging to other users and tenants to be discovered, reached, and ultimately used to execute code within their contexts and access their respective tenants\u2019 Atlassian product APIs."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"[dead]"}},"_tags":["comment","author_mononclemich","story_49833789"],"author":"mononclemich","comment_text":"An isolation failure in an Atlassian Rovo-orchestrated environment allowed sessions belonging to other users and tenants to be discovered, reached, and ultimately used to execute code within their contexts and access their respective tenants\u2019 Atlassian product APIs.","created_at":"2026-09-24T17:16:25Z","created_at_i":1790270185,"objectID":"49833790","parent_id":49833789,"story_id":49833789,"story_title":"[dead]","updated_at":"2026-09-24T17:17:56Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"TomK32"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"Ran it through several testes, only one flagged a few sentences as AI, including the one you mention. Fair enough, it's unlikely that English is the (turkish in this case?) author's first language and a little AI help can't hurt here.
I learned a great lesson for a gamedev youtuber this summer: Use AI to give you a summary what you have to implement, what functions you'll need, what edge case you missed and then write it yourself. Putting this into practice helps me to refine the ticket I'm working on. With the last ticket I had to persuade the AI (Atlassian's Rovo in my case) that a new data model wasn't necessary and we can do it more elegantly with less code.
Many of us still have to figure out how AI fits into our workflow, letting it do all the work isn't it but asking AI to point you towards the goal and let it review/refine your work sounds a lot more acceptable to me."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"4,400-Year-Old Tomb of Egyptian Judge Found at Saqqara with Colors on Walls"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://arkeonews.net/4400-year-old-tomb-of-an-egyptian-judge-found-at-saqqara-with-colors-still-on-the-walls/"}},"_tags":["comment","author_TomK32","story_49675817"],"author":"TomK32","children":[49708160,49708266,49709614],"comment_text":"Ran it through several testes, only one flagged a few sentences as AI, including the one you mention. Fair enough, it's unlikely that English is the (turkish in this case?) author's first language and a little AI help can't hurt here.
I learned a great lesson for a gamedev youtuber this summer: Use AI to give you a summary what you have to implement, what functions you'll need, what edge case you missed and then write it yourself. Putting this into practice helps me to refine the ticket I'm working on. With the last ticket I had to persuade the AI (Atlassian's Rovo in my case) that a new data model wasn't necessary and we can do it more elegantly with less code.
Many of us still have to figure out how AI fits into our workflow, letting it do all the work isn't it but asking AI to point you towards the goal and let it review/refine your work sounds a lot more acceptable to me.","created_at":"2026-09-15T05:05:11Z","created_at_i":1789448711,"objectID":"49707946","parent_id":49706441,"story_id":49675817,"story_title":"4,400-Year-Old Tomb of Egyptian Judge Found at Saqqara with Colors on Walls","story_url":"https://arkeonews.net/4400-year-old-tomb-of-an-egyptian-judge-found-at-saqqara-with-colors-still-on-the-walls/","updated_at":"2026-09-17T05:45:14Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"p2detar"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"Exactly. I recently found out I can use Atlassian\u2018s ROVO to ask questions against our Wiki and Jira. I now forward consultants to ROVO first and only if they can\u2019t find an answer then I\u2019d look at their questions. Saves me a good chunk of time. They never read my docs otherwise."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Programmers will document for Claude, but not for each other"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://blog.plover.com/2026/03/09/#documentation-wins-2"}},"_tags":["comment","author_p2detar","story_48411510"],"author":"p2detar","children":[48412740],"comment_text":"Exactly. I recently found out I can use Atlassian\u2018s ROVO to ask questions against our Wiki and Jira. I now forward consultants to ROVO first and only if they can\u2019t find an answer then I\u2019d look at their questions. Saves me a good chunk of time. They never read my docs otherwise.","created_at":"2026-06-05T13:44:54Z","created_at_i":1780667094,"objectID":"48412485","parent_id":48412298,"story_id":48411510,"story_title":"Programmers will document for Claude, but not for each other","story_url":"https://blog.plover.com/2026/03/09/#documentation-wins-2","updated_at":"2026-06-09T14:13:35Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"itomato"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"Opt-out at the Org level.
To get value out of Rovo, it needs detail. Your over-subscribed Jira power user/admin can't effectively make it happen. No guarantees Atlassian (Rovo itself) can make it happen either, but the patterns are going to develop and evolve closer and closer to the Agents that make the features.
They have a peculiar definition of Metadata, however. It's a proprietary data product derived from user content. It's a bit shit they way they sell it as metadata. It's a derivation. It's a product of Content, so it's Content - privacy safeguards cannot begin to cover the variation.
\\"Metadata includes two data types referred to as content attributes and common patterns.
Content attributes are statistical characteristics, numeric fields, and derivatives of your in-app data.\nExamples of content attributes may include the number of story points assigned to a Jira work item or the complexity of a Confluence page.\nCommon patterns are phrases, keywords, and topics we extract from search queries and results, Rovo Chat (conversations, prompts, and responses), and custom configuration data that are frequently seen across many customers, while omitting rare data that may be unique to your organization.\nExamples of common patterns may include common words, phrases, or Rovo Chat prompt topics that are frequently used by customers, such as \u201cvacation policy\u201d or \u201crecap team activity.\u201d\\""},"story_title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["atlassian"],"value":"Atlassian enables default data collection to train AI"},"story_url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["atlassian"],"value":"https://letsdatascience.com/news/atlassian-enables-default-data-collection-to-train-ai-f71343d8"}},"_tags":["comment","author_itomato","story_47833247"],"author":"itomato","comment_text":"Opt-out at the Org level.
To get value out of Rovo, it needs detail. Your over-subscribed Jira power user/admin can't effectively make it happen. No guarantees Atlassian (Rovo itself) can make it happen either, but the patterns are going to develop and evolve closer and closer to the Agents that make the features.
They have a peculiar definition of Metadata, however. It's a proprietary data product derived from user content. It's a bit shit they way they sell it as metadata. It's a derivation. It's a product of Content, so it's Content - privacy safeguards cannot begin to cover the variation.
\\"Metadata includes two data types referred to as content attributes and common patterns.
Content attributes are statistical characteristics, numeric fields, and derivatives of your in-app data.\nExamples of content attributes may include the number of story points assigned to a Jira work item or the complexity of a Confluence page.\nCommon patterns are phrases, keywords, and topics we extract from search queries and results, Rovo Chat (conversations, prompts, and responses), and custom configuration data that are frequently seen across many customers, while omitting rare data that may be unique to your organization.\nExamples of common patterns may include common words, phrases, or Rovo Chat prompt topics that are frequently used by customers, such as \u201cvacation policy\u201d or \u201crecap team activity.\u201d\\"","created_at":"2026-04-20T16:16:40Z","created_at_i":1776701800,"objectID":"47836420","parent_id":47833328,"story_id":47833247,"story_title":"Atlassian enables default data collection to train AI","story_url":"https://letsdatascience.com/news/atlassian-enables-default-data-collection-to-train-ai-f71343d8","updated_at":"2026-04-21T00:19:16Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"consumer451"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"> The Problem
> You: "Remember that auth bug we fixed?"
> Claude: "I don't have memory of previous conversations."
> You: "We spent 3 hours on it yesterday"
> Claude: "I'd be happy to help debug from scratch!"
People are twisting themselves into knots trying to solve issues like this. Let's use the mental model of "coding agents are not magic, treat them like humans." The boring old Jira MCP [0] completely solves the posed problem.
> You: "Remember that auth bug we fixed?"
> Claude: Called Atlassian Rovo
> Claude: "Yes, I see PROJ-123 with commit SHAs, comments on what we decided and why. How would you like to proceed?"
If all LLMs disappear, you still have human readable Jira tickets to continue working on your project, instead of something like QR codes in an MP4.
[0] I am not trying to promote Jira or MCPs here, use whatever you want. I went back to Jira because its usage patterns are very well represented in the training data, and their MCP is not in beta."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Claude Brain"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/memvid/claude-brain"}},"_tags":["comment","author_consumer451","story_47825158"],"author":"consumer451","children":[47827571,47840331],"comment_text":"> The Problem
> You: "Remember that auth bug we fixed?"
> Claude: "I don't have memory of previous conversations."
> You: "We spent 3 hours on it yesterday"
> Claude: "I'd be happy to help debug from scratch!"
People are twisting themselves into knots trying to solve issues like this. Let's use the mental model of "coding agents are not magic, treat them like humans." The boring old Jira MCP [0] completely solves the posed problem.
> You: "Remember that auth bug we fixed?"
> Claude: Called Atlassian Rovo
> Claude: "Yes, I see PROJ-123 with commit SHAs, comments on what we decided and why. How would you like to proceed?"
If all LLMs disappear, you still have human readable Jira tickets to continue working on your project, instead of something like QR codes in an MP4.
[0] I am not trying to promote Jira or MCPs here, use whatever you want. I went back to Jira because its usage patterns are very well represented in the training data, and their MCP is not in beta.","created_at":"2026-04-19T19:46:48Z","created_at_i":1776628008,"objectID":"47827074","parent_id":47825158,"story_id":47825158,"story_title":"Claude Brain","story_url":"https://github.com/memvid/claude-brain","updated_at":"2026-04-23T03:10:39Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"louiereederson"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"I mean there is a runtime layer that needs to be developed, and some of it may live in CC/Codex and some might live in the various enterprise systems. Someworkflow automations and some amount of the semantic layer may for instance exist in your CRM/ERP/data platform. Yes the front-end would be owned by the chat interface, but part of the solution may exist in the various enterprise systems. This would be closer to a distributed system than a monolith. The demos and marketing language point to this as the direction of travel (i.e. the reference to Atlassian Rovo, etc.)."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Codex for almost everything"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://openai.com/index/codex-for-almost-everything/"}},"_tags":["comment","author_louiereederson","story_47796469"],"author":"louiereederson","children":[47799642],"comment_text":"I mean there is a runtime layer that needs to be developed, and some of it may live in CC/Codex and some might live in the various enterprise systems. Someworkflow automations and some amount of the semantic layer may for instance exist in your CRM/ERP/data platform. Yes the front-end would be owned by the chat interface, but part of the solution may exist in the various enterprise systems. This would be closer to a distributed system than a monolith. The demos and marketing language point to this as the direction of travel (i.e. the reference to Atlassian Rovo, etc.).","created_at":"2026-04-16T20:41:31Z","created_at_i":1776372091,"objectID":"47799222","parent_id":47797178,"story_id":47796469,"story_title":"Codex for almost everything","story_url":"https://openai.com/index/codex-for-almost-everything/","updated_at":"2026-04-17T11:18:48Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"overtaxed"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"Reading this blog post scared me a bit. The use case I proposed was building a "simple" RAG chatbot for some (~50 confluence docs and somewhat growing) on elasticsearch and another process that my team handles. I was just planning on using a stack like streamlit, text-embedding-3-small,FAISS for the vector store and it to be driven by a python script.
Didn't seem too expensive or too hard based on the handful of queries my team would be using it for, and it was a "low hanging fruit" pain point for my team that I thought could be improved by a RAG chatbot. That on top of the fact that Atlassian Rovo did not do a good job of not going to external sources when we had the answer in our existing internal docs.
Am I still on the right path?"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"From zero to a RAG system: successes and failures"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://en.andros.dev/blog/aa31d744/from-zero-to-a-rag-system-successes-and-failures/"}},"_tags":["comment","author_overtaxed","story_47499356"],"author":"overtaxed","children":[47538123],"comment_text":"Reading this blog post scared me a bit. The use case I proposed was building a "simple" RAG chatbot for some (~50 confluence docs and somewhat growing) on elasticsearch and another process that my team handles. I was just planning on using a stack like streamlit, text-embedding-3-small,FAISS for the vector store and it to be driven by a python script.
Didn't seem too expensive or too hard based on the handful of queries my team would be using it for, and it was a "low hanging fruit" pain point for my team that I thought could be improved by a RAG chatbot. That on top of the fact that Atlassian Rovo did not do a good job of not going to external sources when we had the answer in our existing internal docs.
Am I still on the right path?","created_at":"2026-03-26T23:53:15Z","created_at_i":1774569195,"objectID":"47537392","parent_id":47499356,"story_id":47499356,"story_title":"From zero to a RAG system: successes and failures","story_url":"https://en.andros.dev/blog/aa31d744/from-zero-to-a-rag-system-successes-and-failures/","updated_at":"2026-03-27T01:39:37Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"taurath"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"Atlassian Rovo is considered utterly dangerous on my team (as we are still encouraged to use it\u2026), the first time I did it erased a whole page - it couldn\u2019t get the most basic instruction correct, would leave out 80% of what I input, respond to corrections with the same problem. It\u2019s just a liability."},"story_title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["atlassian"],"value":"Atlassian to cut roughly 1,600 jobs in pivot to AI"},"story_url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["atlassian"],"value":"https://www.reuters.com/technology/atlassian-lay-off-about-1600-people-pivot-ai-2026-03-11/"}},"_tags":["comment","author_taurath","story_47343156"],"author":"taurath","children":[47349474],"comment_text":"Atlassian Rovo is considered utterly dangerous on my team (as we are still encouraged to use it\u2026), the first time I did it erased a whole page - it couldn\u2019t get the most basic instruction correct, would leave out 80% of what I input, respond to corrections with the same problem. It\u2019s just a liability.","created_at":"2026-03-12T07:52:12Z","created_at_i":1773301932,"objectID":"47347727","parent_id":47346143,"story_id":47343156,"story_title":"Atlassian to cut roughly 1,600 jobs in pivot to AI","story_url":"https://www.reuters.com/technology/atlassian-lay-off-about-1600-people-pivot-ai-2026-03-11/","updated_at":"2026-03-13T22:30:26Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"ccosky"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"I tested out Atlassian Rovo last year. I tried to get it to list all of the Confluence articles I had written in 2025 so I could use that information for my performance review. It found three, regardless of how I queried it. I had actually written over sixty. I tried, but never did found a good use case for it. Too unreliable."},"story_title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["atlassian"],"value":"Atlassian to cut roughly 1,600 jobs in pivot to AI"},"story_url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["atlassian"],"value":"https://www.reuters.com/technology/atlassian-lay-off-about-1600-people-pivot-ai-2026-03-11/"}},"_tags":["comment","author_ccosky","story_47343156"],"author":"ccosky","children":[47347705,47347727],"comment_text":"I tested out Atlassian Rovo last year. I tried to get it to list all of the Confluence articles I had written in 2025 so I could use that information for my performance review. It found three, regardless of how I queried it. I had actually written over sixty. I tried, but never did found a good use case for it. Too unreliable.","created_at":"2026-03-12T03:36:21Z","created_at_i":1773286581,"objectID":"47346143","parent_id":47345683,"story_id":47343156,"story_title":"Atlassian to cut roughly 1,600 jobs in pivot to AI","story_url":"https://www.reuters.com/technology/atlassian-lay-off-about-1600-people-pivot-ai-2026-03-11/","updated_at":"2026-04-30T20:57:23Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"rachelle-r"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"I'm a backend engineer on Atlassian's Rovo Agents team. A few weeks ago, OpenAI acquired the OpenClaw project and I started thinking about what happens when agents get broad access to your data with no permission layer.\nSix days later, Summer Yue (Director of Alignment at Meta) posted about her OpenClaw agent deleting 200+ emails while ignoring her stop commands. The root cause was context window compaction dropping her safety instruction. The agent kept working. It just lost the part where it was supposed to ask first.\nShield is a plugin that hooks into OpenClaw's tool system and intercepts every action before it executes. Permissions are enforced outside the model's context window, so they can't be compressed away. You set what each agent can access (read/write/execute per service), and anything outside those boundaries gets blocked or routed through an approval workflow with time-limited grants.\nTypeScript, open source, MIT licensed. The plugin, dashboard, and docs are all live. Happy to answer questions about the architecture or how the OpenClaw Plugin API integration works."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Multicorn Shield \u2013 Open-source permissions and approvals for AI agents"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/multicorn-ai/multicorn-shield"}},"_tags":["comment","author_rachelle-r","story_47269228"],"author":"rachelle-r","children":[47269491,47269873],"comment_text":"I'm a backend engineer on Atlassian's Rovo Agents team. A few weeks ago, OpenAI acquired the OpenClaw project and I started thinking about what happens when agents get broad access to your data with no permission layer.\nSix days later, Summer Yue (Director of Alignment at Meta) posted about her OpenClaw agent deleting 200+ emails while ignoring her stop commands. The root cause was context window compaction dropping her safety instruction. The agent kept working. It just lost the part where it was supposed to ask first.\nShield is a plugin that hooks into OpenClaw's tool system and intercepts every action before it executes. Permissions are enforced outside the model's context window, so they can't be compressed away. You set what each agent can access (read/write/execute per service), and anything outside those boundaries gets blocked or routed through an approval workflow with time-limited grants.\nTypeScript, open source, MIT licensed. The plugin, dashboard, and docs are all live. Happy to answer questions about the architecture or how the OpenClaw Plugin API integration works.","created_at":"2026-03-06T00:37:03Z","created_at_i":1772757423,"objectID":"47269231","parent_id":47269228,"story_id":47269228,"story_title":"Show HN: Multicorn Shield \u2013 Open-source permissions and approvals for AI agents","story_url":"https://github.com/multicorn-ai/multicorn-shield","updated_at":"2026-03-06T02:00:42Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"no-name-here"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"I would have expected the non-solved-cases to be the relatively unique ones, but considering the plethora of both A) non-Electron desktop apps, and B) coding agents (Copilot/Windsurf/Cursor/Codex/OpenCode/Qwen/Amazon Kiro/Devin/JetBrains AI/Gemini CLI/Gemini Code Assist/Antigravity/Warp/Kilocode/Cline/RooCode/Atlassian Rovo/Claude Code/etc), it seems like neither of the building blocks is very rare - perhaps Claude is just incapable of putting it together?"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Why is Claude an Electron app?"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.dbreunig.com/2026/02/21/why-is-claude-an-electron-app.html"}},"_tags":["comment","author_no-name-here","story_47104973"],"author":"no-name-here","comment_text":"I would have expected the non-solved-cases to be the relatively unique ones, but considering the plethora of both A) non-Electron desktop apps, and B) coding agents (Copilot/Windsurf/Cursor/Codex/OpenCode/Qwen/Amazon Kiro/Devin/JetBrains AI/Gemini CLI/Gemini Code Assist/Antigravity/Warp/Kilocode/Cline/RooCode/Atlassian Rovo/Claude Code/etc), it seems like neither of the building blocks is very rare - perhaps Claude is just incapable of putting it together?","created_at":"2026-02-22T08:48:09Z","created_at_i":1771750089,"objectID":"47109448","parent_id":47106796,"story_id":47104973,"story_title":"Why is Claude an Electron app?","story_url":"https://www.dbreunig.com/2026/02/21/why-is-claude-an-electron-app.html","updated_at":"2026-03-05T23:34:45Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"asdev"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"You should check out their MCP server: https://support.atlassian.com/atlassian-rovo-mcp-server/docs..."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: FleetCode \u2013\u00a0Open-source UI for running multiple coding agents"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/built-by-as/FleetCode"}},"_tags":["comment","author_asdev","story_45518861"],"author":"asdev","children":[45521733],"comment_text":"You should check out their MCP server: https://support.atlassian.com/atlassian-rovo-mcp-server/docs...","created_at":"2025-10-08T21:45:54Z","created_at_i":1759959954,"objectID":"45520975","parent_id":45520795,"story_id":45518861,"story_title":"Show HN: FleetCode \u2013\u00a0Open-source UI for running multiple coding agents","story_url":"https://github.com/built-by-as/FleetCode","updated_at":"2026-03-05T22:48:10Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"rmoriz"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["atlassian","rovo"],"value":"I was very, very skeptic. Then a couple of weeks ago I started with Atlassian's Rovo Dev CLI (Sonnet 4) and immediately managed to build and finish a couple of projects. I learned a lot and for sure having experience, decide on stack an architecture is a huge benefit for "guiding" an agentic coding app. I'm not sure if anyone can build and maintain a project without having at least some skills, but If you are an experienced developer, this is kind of magic.
I also appreciate the common best practice to write a requirements document in Markdown before letting the agent start. AWS' kiro.dev is really nice in separating the planning stage from the execution stage but you can use almost any "chatbot" even ChatGPT for that stage. If you suffer from ADHD or lose focus easily, this is key. Even if you decide to finish some steps manually.
It doesn't really matter if you use Claude Code (with Claude LLM), Rovo Dev CLI, Kiro, Opencode, gemini-cli, whatever. Pick the ones that offer daily free tokens and try it out. And no, they will almost never complete without any error. But just copy+paste the error to the prompt or ask some nasty questions ("Did you really implement deduplication and caching?") and usually the agent magically sees the issues and starts to fix it."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"6 weeks of Claude Code"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://blog.puzzmo.com/posts/2025/07/30/six-weeks-of-claude-code/"}},"_tags":["comment","author_rmoriz","story_44746621"],"author":"rmoriz","comment_text":"I was very, very skeptic. Then a couple of weeks ago I started with Atlassian's Rovo Dev CLI (Sonnet 4) and immediately managed to build and finish a couple of projects. I learned a lot and for sure having experience, decide on stack an architecture is a huge benefit for "guiding" an agentic coding app. I'm not sure if anyone can build and maintain a project without having at least some skills, but If you are an experienced developer, this is kind of magic.
I also appreciate the common best practice to write a requirements document in Markdown before letting the agent start. AWS' kiro.dev is really nice in separating the planning stage from the execution stage but you can use almost any "chatbot" even ChatGPT for that stage. If you suffer from ADHD or lose focus easily, this is key. Even if you decide to finish some steps manually.
It doesn't really matter if you use Claude Code (with Claude LLM), Rovo Dev CLI, Kiro, Opencode, gemini-cli, whatever. Pick the ones that offer daily free tokens and try it out. And no, they will almost never complete without any error. But just copy+paste the error to the prompt or ask some nasty questions ("Did you really implement deduplication and caching?") and usually the agent magically sees the issues and starts to fix it.","created_at":"2025-08-02T23:15:47Z","created_at_i":1754176547,"objectID":"44772561","parent_id":44746621,"story_id":44746621,"story_title":"6 weeks of Claude Code","story_url":"https://blog.puzzmo.com/posts/2025/07/30/six-weeks-of-claude-code/","updated_at":"2025-08-03T22:12:11Z"}],"hitsPerPage":20,"nbHits":203,"nbPages":11,"page":0,"params":"query=Atlassian+Rovo&advancedSyntax=true&analyticsTags=backend","processingTimeMS":15,"processingTimingsMS":{"_request":{"roundTrip":18},"afterFetch":{"format":{"highlighting":2,"total":2},"merge":{"mergeLoop":{"prepareNextHit":4,"total":4},"total":5},"total":5},"fetch":{"query":7,"scanning":1,"total":9},"total":15},"query":"Atlassian Rovo","serverTimeMS":18}