Not in the field, so genuine question: what is the evidence/theory to support the notion that deep learning is at all a reasonable route towards AGI? As I understand, this is nothing like how actual neurons work - and since they are the only "hardware" which has ever demonstrated general intelligence hoping for AGI from current computational neural networks feel like a stretch, at best.
In short, more suitable staffing and scoping. Hopefully I don't need to explain the arbitrary nature of the 40-hour week.
A lot of people like the idea of flexibility along with these reduced hours, but I'd actually prefer if we went to a uniform 3-day weekend with Friday off. This would have the benefit of aligning the definition of the weekend in the middle East with that in Europe and the US.
I would extend this by saying that most of the people in the US earning mid 6 figures as developers are doing even less good for society than those managers. The market set those salaries high because the negligible marginal cost of reproduction in the software industry is a lure for capital, not because it produces things we actually need.
I agree that LibreOffice is basically a non-starter in this competition, but I also think you are over-selling the MS support experience - most of the information online to help with Word/Excel is rudimentary garbage cluttering up the search space.
We need a set of stand-alone Office alternatives which are fully cross-platform (so, Electron), backwards compatible with Excel spreadsheets and also offering more advanced no-code functionality, and for the Word/PowerPoint equivalents something with the power of LaTeX but an easy to use WYSIWYG fall-back.
From the cloud provider perspective, number 7 is by far the most important but the article really under-sells it.
Cloud providers don't hunt around for SaaS products to rip off; they are approached by customers who are unhappy with the status quo. There could be any number of things the customer dislikes about your VC-floated company - pricing structure, dealing with your sales people, risk of you getting bought by another cloud provider and taken off the market, etc.
Don't blame big companies when your value proposition is wafer thin and your customers are required to be fellow tech bros in order to want to deal with you.
Again, depends heavily on the industry. Also, I'm not entirely sure this is a good thing; I see a lot of "Twitter famous" people fawning over each other at the moment in something resembling a white-collar version of the US tipping culture, but I don't think that kind of behaviour should be allowed to take root more than it already has. One of the other positives we should be taking from a remote-first culture is greater appreciation for the practices of other cultures (since they are now so much more accessible). I hope we're not sliding towards a world where American fake niceness dominates and more reserved modes of interaction are sidelined.
All fine for Big Tech workers, but let's not forget about the large array of others who work in more "pedestrian" parts of the IT industry, ie those working for enterprise shops who are still working through the early stages of cloud adoption.
I think over a longer timescale (~a decade) I would tend to agree. In the short term, there are too many people addicted to face-to-face work environments; often I feel that this is more down to office politics than productivity, but it'll persist for as long as it is effective in giving a career advantage.
Whether working in the office is preferable or even necessary is going to come down to individual teams, not whole companies. The employers who figure that out quickly will do well out of the current labour shortage. Any business insisting on 100% in office or 100% remote is limiting themselves to a subset of the eligible workforce, so it'll be interesting to see how quickly they get out-competed.
A billion dollars today and a billion 10 years ago are not the same thing. It's not just a question of inflation either; there is more money sloshing around in VC (or similar non-retail funding sources) than ever.
I think the whole idea is idiotic, but this particular point doesn't seem like it would be the blocker. We already have employment contracts which state hours to be worked and similar mundane details, this wouldn't be any different. Such contract details would only be tested if someone tried to exploit a loophole, like showing up for a bunch of interviews to get money but with no intention of actually completing the tasks.
Alternatively, stop pretending this industry is a precious snowflake and conduct reasonable-duration, non-exploitative interviews like everyone else in the history of modern employment.
HPL is a stress test with a notionally useful output measurement - just about the most effective way of pushing CPU load to the limit, and it tells you what fraction of the theoretical maximum FLOPS you can actually achieve given the other system constraints like memory and network throughput.
Employees outside the games division already had much slacker rules about game development; I expect this change aligns things for all AWS employees at the lower bar.
Is it better to aspire for balance by chasing both, or neither? After all, what is status if you have no-one to brag to - and what good is having a lot of free time when you are just going to end up dead either way, and with work being the best chance most of us get at making a mark on the world.
Not meant as a snide/flippant remark; this shit is what keeps me up at night.
Highly debatable. While Amazon is more likely to be on the shit-list of Joe Public, in technical circles it comes down to AWS vs MS. Frankly if MS hadn't bought GitHub to distract magpie developers, there would be no contest.
The part I find baffling is how companies like Spotify justify paying high salaries for engineers just to deliver trivial improvements over what Windows Media Player was offering 20 years ago. If they were piling this effort into reducing costs or making the app more resource efficient I'd be happy - but I'm not interested in trivial UI tweaks.
The rebate you mention sort of already exists, just restricted to public sector research customers. Think the cap is that egress is waived as long as it is less than 15% of the bill.
We could start with something similar to the concept of Pareto optimality; Intel could have delivered their maximum performance without preventing optimizations from being applied equally on AMD hardware, but instead they choose to disadvantage AMD without providing anything extra on top of what they could do while remaining "neutral".
This is trivially easy to defeat, just so you know. If anyone reading is ever in need of optimized math library performance on AMD, just speak to your hardware/cloud vendor; they all know the tricks.
"Should be" is the operative part of your statement - I totally agree there, and that the salesmanship requirement indicates a systemic failure. I'd say it goes much further than just the educational institution being at fault - they are responsible for playing the game, but funding sources do nothing to drive them more towards academic integrity.
I was described as "too scholarly" by my PhD advisor when writing up my first couple of papers. I'm in industry now.
People doing a PhD/post-doc need to understand that there is no escape from "sales". Lots of academically leaning people, including myself, set out with a goal of avoiding becoming a seller and wanting to work in a space where data does the persuading rather than smooth talk. Unfortunately no such place exists. Even in the "purest" science-focused workplaces, you still need to sell your ideas to managers or funding agencies. To transcend "regular researcher" and become widely respected in your field, having thousands of Twitter followers will be more helpful than a paper in Nature.
It's no surprise, that was classic MS strategy. Sell the thing everyone actually needs (Office, because they smothered the competition), but bundle it with something sticky which will also hoover up data about your workers - no need for the product to be any good, execs in charge of purchasing won't have to suffer with it for a fraction of the time employees spend. Oh also, I happen to have $200k of Azure credits here for you - no pressure but if you do start using Azure we'll make the bill for all this productivity software go away...