|
Proofreading, or in current speak, Pull Request reviews
if over 100 lines, quick glance and accept
1 line change, 20 comments about how to do it differently
Even better, is have tests written by the AI for it's own tests.
"do the tests pass?"
ai: "yes 🙄"
code change accepted
2 months later - wait all this test does is Assert.true() at the end, regardless if works.
|
|
|
|
|
maze3 wrote: if over 100 lines, quick glance and accept
1 line change, 20 comments about how to do it differently
No truer words have I read this week. 
|
|
|
|
|
"Improved AI models and techniques have also increased the level of acceptance of suggested code, from 27 percent in June 2022, to 46 pecent today, and 61 percent for Java code." (from the "Github claims" article of the newsletter)
If that sentence had been a micrometer closer it would have mortally struck at the heart of a major "AI" issue.
This presumption that because it gains first-shot acceptance, then it must be pretty good is totally bonkers.
It springs from ignorance but all the same... Code is "young". It's not like banking, or retail, or prostitution. Well, maybe sometimes it's a little like all of those.
The point is that it's not been around so long and the demand exploded so fast. It might be reasonable to say that the average programmer is really not so great at it just by virtue of comparison to more mature industry. So then, it is foolhardy to suggest that more of us passing along the AI's code without alteration means that the AI's code is even necessarily improved.
I get their point. As good at their craft comparably as the average blacksmith? Maybe not, but it is the average we have. Right now. Which is my point.
It knows what we've taught it which was only ever as collectively good as we were. Even if you were to generously tack on a fair bit of "greater than the sum of its parts", you'd still come up short because that is inherently assuming we'd never collectively be significantly better.
Which may be totally correct... if we're ever largely rubbing stamping AI code. It's the same as the art, music, and other things. I'm not really so much against AI or anything of the sort, I've dug it all quite a bit.
It's just the inescapable knowledge that sure, it's all well and good until it's just the AI copying itself and it devolves back to nothing but noise and leaves a bunch of blubbering meat sacks who kind of understand they need food.
|
|
|
|
|
Valhalla is nothing short of a Java language overhaul, promising to correct longstanding performance issues. Here's a first look at what's coming, starting with the new value classes and primitive types. I'm not sure naming your project after a place where dead warriors go to drink and fight is a good idea
|
|
|
|
|
Would you prefer Ragnarok?
I’ve given up trying to be calm. However, I am open to feeling slightly less agitated.
|
|
|
|
|
For Java? Absolutely!  
TTFN - Kent
|
|
|
|
|
Quote: Java ... remains at the forefront of programming language design ...
... by finally implementing a feature that .NET has had for over 20 years.
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
A scathing blog post details dysfunctional life inside Google. "To serve man"
I'd say they still serve users - directly to the advertisers.
|
|
|
|
|
As generative language models become more accessible, easier to scale, and write more persuasive text, they could be used to spread disinformation. I guess they have all the chips they need then?
|
|
|
|
|
Kent Sharkey wrote: ...they could be used to spread disinformation.
With the current torrent of disinformation, the addition would probably go unnoticed! 
|
|
|
|
|
Define disinformation without including a political reference.
|
|
|
|
|
Intentionally using incorrect information to sway opinions.
Depending on how loose you want to apply the political label, I’d include Tokyo Rose et al, and the efforts by the tobacco companies and lobbyists.
TTFN - Kent
|
|
|
|
|
Kent Sharkey wrote: I’d include Tokyo Rose et al, and the efforts by the tobacco companies and lobbyists
Only? Do we live in the same world?
M.D.V. 
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
I wanted to keep it only to a few examples. Otherwise, we’d be here all week.  
TTFN - Kent
|
|
|
|
|
obermd wrote: Define disinformation
Any information that conflicts with my prejudices. 
Freedom is the freedom to say that two plus two make four. If that is granted, all else follows.
-- 6079 Smith W.
|
|
|
|
|
The optimization mode requires quantum effects, can solve a growing list of problems. And a dessert topping
|
|
|
|
|
In future Windows 11 updates, the whole process of killing a process will be much easier than it is today. A small hammer attached to the computer?
|
|
|
|
|
So they are working to make "Killing processes" easier... why don't they better work on how to avoid having to kill processes (specially, the windows internal processes)?
M.D.V.
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
Being able to kill tasks a user launches without having admin access to the system, which is required for Task Manager, is a huge boon to overall system security.
|
|
|
|
|
Microsoft wants to enhance Windows with smarter features. Because that's what AI is for - moving windows on the screen
|
|
|
|
|
Kent Sharkey wrote: moving windows on the screen
They have probably gone out of new icons... now is their position in the mosaic of the monitor in the row
M.D.V.
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
GitHub has updated its Copilot coding assistant with new features including a “vulnerability filtering system” to block insecure coding patterns such as SQL injection or hard-coded credentials. No insecure code, just insecure developers
|
|
|
|
|
So am I reading this right - 40-60% of developer output is insecure code? 
|
|
|
|
|
I can introduce you people where that actually rises to 80-90%  
M.D.V.
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
The eye-waveringly expensive bid by Adobe was made public in September and upset the applecart for many creative types that are concerned Adobe will merge its new toy with its XD software and raise prices.
Graeme
"I fear not the man who has practiced ten thousand kicks one time, but I fear the man that has practiced one kick ten thousand times!" - Bruce Lee
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
