Forgot your password?
typodupeerror

Comment: Re:Only if the PC is made by Apple (Score 1) 60

by biodata (#47414091) Attached to: All Web Developers Should Have Access to a Device Lab (Video)
I don't think people are bothering to check Safari compatibility any more, unless they happen to use a Mac themselves. I'm finding more and more sites where Safari has buttons completely missed from the rendering (or hidden behind something else or otherwise invisible). If Safair won't run on anything but a mac then a large proportion of the dev community is never going to bother checking it.

Comment: Censorship through comment (Score 3, Informative) 74

by biodata (#47385377) Attached to: Google Reinstating Some 'Forgotten' Links
The Preston case was particularly pernicious - a whole article disappears from search results just because one person adds a comment to the article then decides to 'retract' their comment because 'it is not relevant any more'. It would have introduced a very easy attack route for anyone to take down any article they didn't like by posting a comment then asking Google to retract it thus hiding the whole article.

Comment: Re:E-mail? (Score 1) 346

by biodata (#47378525) Attached to: Goldman Sachs Demands Google Unsend One of Its E-mails
This x 10^6 Anyone who has used the internet should have realised pretty early on that sending an email is equivalent to sending a postcard - it will probably arrive at the destination if you use the right address, but there are no guarantees, and you have no control over who sees the email on the way as it is not a private protocol. People should have to get a licence by demonstrating minimal competence before they are allowed to use potentially dangerous technology, just as they do with driving a car.

Comment: Average across models (Score 5, Informative) 230

by biodata (#47099025) Attached to: The Flaw Lurking In Every Deep Neural Net
Neural networks are only one way to build machine learning classifiers. Everything we've learnt about machine learning tells us not to rely on a single method/methodology and that we will consistently get better results by taking the consensus of multiple methods. We just need to make sure that a majority of the other methods we use have different blind spots to the ones the neural networks have.

No amount of careful planning will ever replace dumb luck.

Working...