Timeline for Our partnership with Google and commitment to socially responsible AI
Current License: CC BY-SA 4.0
20 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 9 at 20:53 | history | edited | leftaroundabout | CC BY-SA 4.0 |
added 1370 characters in body
|
| Mar 14 at 9:56 | comment | added | leftaroundabout | @TravisJ that's certainly something I'm concerned about too, but we should fear it from OpenAI etc. rather than from StackOverflow. SO's asset is specifically the human contributors and they'd be mad (not just in an idealistic-mission sense, also from a pure business perspective) to forgo that by letting such a feedback loop destroy the advantages it offers them over the pure AI companies. | |
| Mar 14 at 5:22 | comment | added | Travis J | You left out the part where the Ai begins to self train on its own output in all of this and takes a complete nose dive. What do you think is going to happen when clueless consumers begin to regurgitate remixed prompt outputs back into the space that Ai trains from? It simply is not accurate enough to dogfood its own output. That said, if proper attribution were used... at least it could know not to eat its own crap. | |
| Mar 1 at 13:51 | comment | added | President James K. Polk | The company owners expect to be billionaires or it's a disaster. The people who curate the site are volunteers. There is a lot of room between those two for a non-profit site that hews to the original vision of Stack Overflow without compromise. Attribution always has been and always will be an issue that needs attention. As an aside, it's interesting that when I asked Gemini a programming question, its answer included attribution. | |
| Mar 1 at 8:11 | comment | added | leftaroundabout | Apart from that - well, I consider the current style of AI with their lack for attribution considerations or human peer-review ethically problematic, and even if they offered to all help-seekers more satisfying solutions than SO does then this is not a future we should welcome. Again, I'm not convinced that what Google cooks up here will be ethical either, but I also can't see any better suggestions for how to take steps in the right direction. | |
| Mar 1 at 8:02 | comment | added | leftaroundabout | @PresidentJamesK.Polk false dichotomy. We may neither have superior AI just around the corner, nor any guarantee that SO stays relevant because it's still the better option. Instead, what we have is AI that's good enough - and more convenient - for so many users (95%? 99%?) that it reduces SO to a niche site with too little traffic to really stay up to date, while we more experienced users still need it as a more robust platform / resource compared to the often-erratic AI alternatives. | |
| Mar 1 at 6:52 | history | edited | This_is_NOT_a_forum | CC BY-SA 4.0 |
Active reading [<https://en.wiktionary.org/wiki/milk#Verb> <https://stackoverflow.design/brand/copywriting/naming/>]. Dressed the naked link.
|
| Feb 29 at 23:45 | comment | added | President James K. Polk | As a result, SO would end up being more or less a ghost town plundered for its past wealth but with little relevant new activity. Only if the AI tools are better. And if they are better then why should I care that SO is a ghost town? This answer shares a view with many of the angry delete-my-account rants in that it assumes that a world of superior AI is just around the corner and that, therefore, we should strive to keep SO relevant. I don't see that world just around the corner, but if it comes and obsoletes SO then why should I care? I don't own shares in the company. | |
| Feb 29 at 18:02 | comment | added | Kevin B | I don't feel it's worth the effort. They aren't listening anyway. | |
| Feb 29 at 17:57 | comment | added | leftaroundabout | @KevinB fair enough, why don't you elaborate that in an answer? And also, what would you suggest SO should have done differently to not end up where we are now? | |
| Feb 29 at 17:55 | comment | added | Kevin B | That's exactly our criticism. they have a stance that their actions aren't supporting. | |
| Feb 29 at 17:54 | comment | added | leftaroundabout | Well, clearly the three of them can't be all non-hypothetical simultaneously, other than in a Schrödinger's Cat way... | |
| Feb 29 at 17:52 | comment | added | Kevin B | My point is your hypothetical options aren't hypothetical. They've happened, and are very real possibilities for things that are also possibly being launched in the near future. We're way past #2 being a reality at this point, the entire company has been reorganized toward the purpose of building and selling AI solutions. | |
| Feb 29 at 17:50 | history | edited | leftaroundabout | CC BY-SA 4.0 |
deleted 2 characters in body
|
| Feb 29 at 17:41 | comment | added | leftaroundabout | I somewhat agree that 3 already occured, but that content will get stale. An AI that just bases its help on SO posts from 2022 will be pretty useless for many programming problems in 2027. | |
| Feb 29 at 17:38 | history | edited | leftaroundabout | CC BY-SA 4.0 |
added 33 characters in body
|
| Feb 29 at 17:36 | comment | added | Kevin B | 3 already occured. No amount of SO blocking it in the future will undo the fact that all of the best content has already been taken. This partnership literally gives all current and future content generated by the community to google to be used for training, not just for OverflowAI, but for whatever they want to do with it. They're certainly not going to start citing questions from SO in their code completion features. | |
| Feb 29 at 17:36 | comment | added | leftaroundabout | @KevinB 1.-3. are hypothetical alternative realities. I'll clarify the post. | |
| Feb 29 at 17:35 | comment | added | Kevin B | I don't trust SE on this statement: "SO takes a stance to keep the platform itself AI-free but the data fully available for anybody." given they've literally experimented publicly with the opposite of this. | |
| Feb 29 at 17:29 | history | answered | leftaroundabout | CC BY-SA 4.0 |