Do a piece on GoDaddy - they're absolutely disastrous. Eg, when you need to do any kind of downgrade of a service, you need a 2fa over email which they intentionally delay for like 15mins. If you leave the page, you have to request a new one.
Of course, when it's 2fa for something that's not negative in terms of revenue, it arrives immediately.
Another fun one for godaddy: 'forgetting' to email you about a renewal, letting it lapse and then charging you literally double to "reactivate it" when the DNS stops working.
Happened to my company, I moved DNS to porkbun immediately after reactivating the domain. Wasn't a cheap one either so that cost was especially unwelcome.
Do they still do the thing where if you search for a domain name and decided not to buy it they mysterious buy it themselves a couple days later and increase the price threefold?
GoDaddy flagrantly steps over the line from deceptive practices into unethical shady business territory. Continues to amaze how they are still running.
To this day I am not entirely convinced that they did not start as something sex-industry-related and then for some strange reason switched to DNS registrar business later. Seriously, "Go, daddy!"?
Despise Godaddy. And google domains sold to Squarespace, who makes a transfer take up to 15 days. It’s faster to carve the document from wood and send it across the world than it is for them to automatically click the button to let go of a domain. I’ve never had a transfer go that slowly anywhere.
Gemini Ultra should show good progress according to Google - it's supposed to perform better than 85% of computer science competitors, which requires a lot of logical reasoning.
Lets see it once it goes live, but sounds promising.
Their previous model was better than 46% of such competitors (according to them), so 85% seems achievable by throwing more compute resources at typical ML training. After all, training on millions of examples of logical reasoning will undoubtedly store logical rules in the model in some shape or form (it does so even in ChatGPT), yet the results are still more "convincing" rather than "correct", or "probably correct" at best, usually achieved with lots of postprocessing on top. GPT-4 is better than 90% of lawers at the bar exam, yet still manages to fail at reasoning on much simpler domains.
It's a great list. Would be a good to keep it as some live list so we can help evolve it over time.
I'd add two ideas:
Build lightweight memcached
- it requires mixture of basic algorithms as well as good system design practices
- it's a "simple" problem, but requires to think well from multiple angles - invalidation, concurrency, memory management...
Build your own docker
- it's not as complex as it might sound
- gives opportunity to understand basics of OS programming, which most engineers nowadays don't really use in day to day jobs
The muon g-2 measurements have nothing to do with floating rocks. This is not a superconductor update.
This is a high energy physics thing about failing to find something beyond the standard model. Just your regularly scheduled "5 sigma physics anomalies that disappear on their own with better measurements" programming.
Right. I mean that measurements conflicted with theory, and the new CMD-3 measurements suggest a better calculation. Assuming CMD-3 is good, there's no significant gap between theory results and collider results
That wasn't the only change. There was an alternate computational scheme that didn't depend on electron-positron experiments that gave results consistent with the Muon g-2 experiment.
Would be smart to release windows phone support for free, but IOS and android support costs. More windows phone apps for app store for no cost to developer.
