Copilot regurgitating Quake code, including sweary comments • Hacker News Board

Re: Copilot regurgitating Quake code, including sweary comments

#181

Post by buu700 » Fri, Jul 02, 2021, 2:18 PM UTC

gpm wrote: post #160

buu700 wrote: post #146

Earlier quoted context omitted.

Could they handle this by generating a collective attribution file that covers every (permissively licensed) repository that Copilot learned from? Of course this would be massive, so from a practical consideration the attribution file that Copilot generates in the local repository would have to just link to the full file, but I don't think that would be an issue in and of itself.

Maybe? Might depend on the license, I doubt the courts would be amused. Almost certainly a link would not suffice, basically every license requires that the attribution be directly included with the modified material. Links can rot, can be inaccessible if you don't have internet access, can change out from underneath you, etc. (I am not a lawyer, btw)

Makes sense. Maybe something like git-lfs/git-annex would be sufficient to address the linking issue, but it seems like the bigger concern is whether a court would accept this as valid attribution. In a sense it reminds me of the LavaBit stunt with the printed key.

Re: Copilot regurgitating Quake code, including sweary comments

#182

Post by bidirectional » Fri, Jul 02, 2021, 2:18 PM UTC

jhugo wrote: post #175

bidirectional wrote: post #163

Earlier quoted context omitted.

Every front office finance project I have ever worked on has used floating point, so take the dogma with a grain of salt. It depends entirely on the context.

They probably just accumulate the rounding errors into an account and write it off periodically without even realising why it happens.

No, it's just that we're in the realm of predictions and modelling, not accounting. If you're constructing a curve to forecast 50 years of interest rates from a limited set of instruments, you're already accepting a margin of error orders of magnitude greater than the inaccuracies introduced by floating point.

The models also use transcendental functions which cannot be accurately calculated with fixed point, rationals, integers etc.

Re: Copilot regurgitating Quake code, including sweary comments

#183

Post by dempsey » Fri, Jul 02, 2021, 2:18 PM UTC

I’ve always wondered this about the realistic photo generators. How do we know they’re generating new faces and not just regurgitating ingested faces?

Re: Copilot regurgitating Quake code, including sweary comments

#184

Post by School-Cotton » Fri, Jul 02, 2021, 2:19 PM UTC

jacobsenscott wrote: post #154

School-Cotton wrote: post #110

Earlier quoted context omitted.

Depends what you’re doing. In fact it’s not always wrong to use floats for currency. For accounting you should probably use a fixed-precision decimal type.

If someone asks how to handle money the best answer is integers or fixed precision decimals. There may be a valid case for using floats, but if someone asks they shouldn't be using floats. Also I'm hard pressed to come up with a case where floats would work. Can you give an example?

> Can you give an example?

The answer is the same as _any_ time you should use floats: where you don't care about answers being exact, either (1) because calculation speed is more important than exactness, or (2) because your inputs or computations involve uncertainty anyway, so it doesn't matter.

This is more likely to be the case in, say, physics than it is in finance, but it's not impossible in the latter. For example, if you are a hedge fund and some model computes "the true price of this financial instrument is 214.55", you certainly want to buy if it's being sold for 200, and certainly don't if it's being sold for 250, but if it's being sold for 214.54, the correct interpretation is that _you aren't sure_.

When people say "you should never use floats for currency", their error is in thinking that the only applications for currency are in accounting, billing, and so on. In those applications, one should indeed use a decimal type, because we do care about the rounding behavior exactly matching human customs.

Re: Copilot regurgitating Quake code, including sweary comments

#185

Post by tyingq » Fri, Jul 02, 2021, 2:20 PM UTC

slver wrote: post #87

tyingq wrote: post #30

They have 4 hand picked examples on their homepage: https://copilot.github.com/ One has the issue with form encoding: https://news.ycombinator.com/item?id=27697884 The python example is using floats for currency, in an expense tracking context. The golang one uses a word ("value") for a field name that's been a reserved word since SQL-1999. It will work in popular open source SQL databases, but I believe it would bom…

> And these are the hand picked examples. This product seems like it needs some more thought. Everyone's self-preservation instincts kicking in to attack Copilot is kinda amusing to watch. Copilot is not supposed to produce excellent code. It's not even supposed to produce final code, period. It produces suggestions to speed you up, and it's on you to weed out stupid shit, which is INEVITABLE. As a side note, Excel a…

"self-preservation"

My suggestion was a way to comment or flag, not to kill the product. These were particularly notable to me because someone hand-picked these 4 to be the front page examples of what a good product it was.

Re: Copilot regurgitating Quake code, including sweary comments

#186

Post by fredros » Fri, Jul 02, 2021, 2:21 PM UTC

ainar-g wrote: post #122

ronnier wrote: post #104

Earlier quoted context omitted.

In its base unit. So cents in USD. Which can be an int64 Or if your language has something specific built in, use that.

> Or if your language has something specific built in, use that. Unless your language is PostgreSQL's dialect of SQL, apparently. https://wiki.postgresql.org/wiki/Don%27t_Do_This#Don.27t_use...

Of course for databases you should use a decimal.

Re: Copilot regurgitating Quake code, including sweary comments

#187

Post by hawski » Fri, Jul 02, 2021, 2:22 PM UTC

Copilot may do more to move open source projects out of GitHub than the message that Microsoft is the buyer. Now you can host the code on GitHub to get your license violated, or DMCA-ed in a long run, when your code will become a part of some big proprietary project. At least it makes me think about my choice for code hosting more then whatever happened before.

Re: Copilot regurgitating Quake code, including sweary comments

#188

Post by hallqv » Fri, Jul 02, 2021, 2:23 PM UTC

louthy wrote: post #5

This is utterly damning. I have already instructed my team that Copilot can never be used for our projects. Compromising the product because of unknowable license demands isn't acceptable in the professional world of software engineering. But if we put the licensing to one side for a moment... 1/ Everything I've seen it generate so far is 'imperative hell'. It is practically a 'boilerplate generator'. That might be u…

Neural net: ”It’s all in the training data, stupid.”

Re: Copilot regurgitating Quake code, including sweary comments

#189

Post by unknownOrigin » Fri, Jul 02, 2021, 2:24 PM UTC

I'm honestly kinda amazed this as upvoted here as it is. Typically anything ML-related is upvoted to the top positions and any dissent harshly ridiculed. Anyways... it appears those who thought about this as if it was a glorified code search engine were close to being right.

Re: Copilot regurgitating Quake code, including sweary comments

#190

Post by ocdtrekkie » Fri, Jul 02, 2021, 2:24 PM UTC

Probably an excellent reminder that both Google and Microsoft decided to use your private emails for a training set to create Smart Reply behavior that can "write emails for you", and they swore up and down there's no way that could ever leak private information.

We need legislation banning companies from ingesting data into AI training sets without explicit permission.