We run a lot of varied, tiny, simple workflows that were previously running on 5.4-nano and mini. We transitioned them to 5.6 and noticed exactly this range of improvement across the board. In a few cases, we had improvements in classification.
I think a lot of people miss that for many companies, a model upgrade like this is basically a one liner.
Even if you have an amazing model router architecture (which we do for our golden flows), it’s just not worth it. Not to mention reliability and so on
People usually think of "making it" as a single point in time that's binary. Granted, there's some truth to that: most people will feel fully comfortable with a few millions in their bank account.
But in reality, success can be savored a lot more pleasantly if you allow yourself to be conscientious of milestones and include having a great process of getting there as the definition of success.
I've met so many people (myself included) that burned themselves by thinking they will spend only a few years grinding to reach a certain milestone, and get frustrated when the narrow definition of success they've created for themselves isn't met.
I've also met a few others that have realized the above, usually with burnout then self-discovery, many which also "made it" by the standard definition, but have a strong sense of fulfillment.
Naturally the question is how. For myself, what worked is a mix of taking more risks (becoming a founder at 35+), experimentation and using various techniques (therapy, meditation, the usual suspects) to find what can drive me to work hard but with purpose or joy.
We’ve been using Microsoft Teams as well as the entire office suite, and we’ve been positively surprised. There is an occasional clunky UI you come across, but the feature set is far superior to Slack or Zoom, and the ecosystem integration is nice.
msft had a massive edge. it had exclusive access to models + had web search before anyone.
they flopped this royally, just like windows mobile. they created a shitty ux by shoving it inside the bing app, then they decided to charge for it instead of capturing all enterprise value.
lastly, the product has stalled and missed on their biggest opportunity which is tapping into the data. you can think it's because of how complex it must be, but then openai and everybody else did it.
it's truly a lesson in product mismanagement, once again, from microsoft
The downstream issues are not related to immigration, otherwise house prices would have fallen in Lisbon.
Look at examples such as Singapore, Dubai and many others that adopted an entrepreneurial attitude and figured out the infrastructure to support that growth.
When will they blame when the immigrants don’t come but the problems remain?
What I see in comment sections of articles like this is a perfect example of why EU is set to lose in the coming decades: mindset.
I’d say most Americans understand the value of highly skilled migration. It’s how America stays #1, it’s how they have the best companies, and bring the most value to their nation.
Meanwhile in EU, any move to attract talent is seen as net-negative, unfair and detrimental to the culture and livability. Their negativity is a self fulfilling prophecy, but it will be a very costly and hard pill to swallow, just like the UK is seeing after Brexit.
I created my own translation app using llama3-80b, I call it "expat translator": I live outside of my home country and always struggled with using translators like Google Translate because they don't tell you if the way you're writing something feels natural in the other language. It gives me some pretty good results and I also instruct it to give me rewrites for informal and professional use, so I don't sound weird on WhatsApp for example.
It uses an on-device model for language detection and results are sub 0.3s thanks to groq
I started using Edge due to work being on the msft suite and I’ve been positively impressed with its performance. It handles way better than Chrome, especially with multiple tabs.
It's a fiscal year convention. Companies use the following year to reference the fiscal year because the fiscal year is named after the year in which it ends.
I have a type of migraine called vestibular migraine and I can confidently say that using dark mode helps A LOT. I agree with previous comments, dark mode is not about optimising for the median, it's about giving more options, just like we did with screen readers, font sizes, etc.
It's one of the sad sides of tech: we have all this knowledge about accessibility and this is either siloed or just not talked about. Designers are learning about aesthetics, color theory, typography but not about making your product support as many humans as possible - which is a business outcome!
Mac and iPhone prices will always be about perceived value. Unfortunately they don't have enough competition to drive prices down and they will take advantage of that for as long as they can.
Some tried from different angles but it seems people just don't feel confident enough to leave the ecosystem. Google keeps pushing sub-par phones. No one has delivered a robust face recognition tech as good as Face ID. Screen and camera quality are almost there but never better in a distinctive manner.
In fact, no one is trying to compete with Apple anymore. Android manufacturers abandoned the high-end market. The tablet and smartwatch markets are pretty much dominated by iPad and Apple Watch.
And as the service category expansion shows, people are getting comfortable with the idea of paying a monthly price for good tech. If competition eventually becomes a problem, they can just start bundling things up like Amazon does: iPhone Upgrade Plan + Apple Music + Cloud Storage + Apple Care + (eventual) TV content.
I think this is a common mistake developers make when creating products: when things start not working, blindly start to add features or improvements they think will be good. I've been on that path myself and had the pain of (almost) going bankrupt.
Most (if not all) of this can be avoided if you spend time after development on getting feedback from customers, doing interviews, usability testing and most importantly trying to find the right commercial angle to develop the business. For example:
- Is the game localised in different languages?
- Why targeting Linux and Windows, and not testing/developing on Mac or mobile platforms?
- Why not create a end-of-year sale or bundle the game with others for promotion?
This is a misleading way to compare salaries and cost of life between countries.
A much better metric is to look at net disposable income after tax, if you're focusing on the "making money" aspect.
If you're focusing on the life quality aspect you'll also have to consider mortgage, car ownership and time off.
Example:
- In Netherlands you have 24 business days of time off, where you can take them from the moment your contract begins (not after 1 year working), and you can take them however you want.
- Taking a mortgage in Netherlands is currently cheaper than renting and most houses in the "sweet spot" cost around 300k euros, which is ridiculously cheap. Monthly payments are subsidized by the government (for some years) which makes it cheaper than renting.
To summarise, every country has a lot to offer, and at the end of the day life is not just about money.
I think a lot of people miss that for many companies, a model upgrade like this is basically a one liner.
Even if you have an amazing model router architecture (which we do for our golden flows), it’s just not worth it. Not to mention reliability and so on