I have used Github Copilot extensively within VS Code for several months. The autocomplete - fast and often surprisingly accurate - is very useful. My only complaint is when writing comments, I find the completions distracting to my thought process.
I tried Gemini Code Assist and it was so bad by comparison that I turned it off within literally minutes. Too slow and inaccurate.
I also tried Codestral via the Continue extension and found it also to be slower and less useful than Copilot.
So I still haven't found anything better for completion than Copilot. I find long completions, e.g. writing complete functions, less useful in general, and get the most benefit from short, fast, accurate completions that save me typing, without trying to go too far in terms of predicting what I'm going to write next. Fast is the key - I'm a 185 wpm on Monkeytype, so the completion had better be super low latency otherwise I'll already have typed what I want by the time the suggestion appears. Copilot wins on the speed front by far.
I've also tried pretty much everything out there for writing algorithms and doing larger code refactorings, and answering questions, and find myself using Continue with Claude Sonnet, or just Sonnet or o1-preview via their native web interfaces, most of the time.
I see, perhaps with Gemini because the model is larger it takes longer to generate the completions. I would expect with a larger model it would perform better on larger codebases. It sounds like for you, it's faster to work on a smaller model with shorter more accurate completions rather than letting the model guess what you're trying to write.
For me it starts with a small patch of my visual field going blind - not white or black, just not able to see anything or make out detail in that area. Typically it's a spot in the center of my vision such that I can't read. After 5-10 minutes, flashing, swirling colored patterns will appear in that same spot. The affected area then enlarges progressively over the course of anywhere between 30 minutes and several hours, covering a polygonally shaped area with the flashing swirling patterns. Sometimes it will stop after 2-3 hours and gradually disappear. Other times the area of blindness and flashing colors will continue to expand for 6 or more hours until it covers an entire half of my visual field.
After the visual aura disappears, usually (but not always) an intense headache will start that lasts the rest of the day.
Triggers for me are unknown but in the past have included air pollution (e.g. smoke from wildfires) and very bright light (watching a video with a white background for too long on a high brightness TV).
I have also noticed greatly reduced frequency of these migraines after I started taking magnesium glycinate daily.
Atlassian IPO'd in 2015... so you would have worked there 2006-2009. Congrats on the success, but that was more than a decade and a half ago... I'm not sure it counts as a recent example.
I'm in this category, except became an invisible solopreneur after deciding startups are a raw deal instead of going back to FAANG. Between the unfavorable tax treatment, long time to IPO, the inherently high risk of options, it's just not worth it. Perhaps a good deal for the VCs, maybe for the founders, but not for even the earliest employees. See you in the next life.
It has worked just fine under WSL for many months now, including full GPU support, though that's not as convenient for most. Native Windows support is icing on the cake.
Indeed, WSL has surprisingly good GPU passthrough and AVX instruction support, which makes running models fast albeit the virtualization layer. WSL comes with it's own setup steps and performance considerations (not to mention quite a few folks are still using WSL 1 in their workflow), and so a lot of folks asked for a pre-built Windows version that runs natively!
Going out on your own may not be always strictly better or more rewarding, but this founder story is just one datapoint. Since most companies have an employee equity pool that is sized at only 10-20%, working for them is by definition giving 80-90% of the value you create away to the founders and investors. The rewards of either starting a company or being a sole proprietor (e.g. indie dev or consultant), can therefore be expected to be very substantial, on average, perhaps even 5-10x per unit time worked, and that's not including the secondary non-financial benefits of not working for someone else (e.g. freedom, autonomy, etc.). I do wish more people were aware of this principle, because the willingness of engineers in the aggregate to acquiesce to 10-20% ownership is why VCs and founders become so inappropriately rich by comparison. Just my 2c.
Especially with the widespread availability of vapes, which are still pretty bad for you but substantially less so than cigarettes, there's no excuse these days.
I am the same way, I will risk getting hit by cars sometimes just to avoid smoke I hate it so much. Even outdoors after a brief exposure it gets lodged in your hair, clothes, and eyes. With the legalization of marijuana in so many places, the problem has become even worse.
That was from bootleg THC carts that had adulterated ingredients [0]. In this context we're talking about vaping (nicotine liquid) vs smoking (cigarettes).
Vaping is not as safe as inhaling nothing but air. But we do know scientifically that it is substantially safer than smoking [1], despite the (big tobacco-funded) scare campaigns that claim otherwise.
In other words, nothing > vaping > smoking. If you don't use nicotine today, don't start. But if you want to get off cigarettes, vaping is a great choice and provably safer.
Some, but not all [1]: "... the remaining 17% reported using only nicotine-containing vaping products... It is believed that the non-THC/CBD vapers who were diagnosed with EVALI were actually suffering from disparate vaping-associated lung diseases caused by different toxins within the aerosols or different host responses to the inhalants."
The CDC recommends not using THC vapes at all because of EVALI, but make no such recommendation about standard nicotine vapes (unless you've never smoked) [0].
Totally agree, it definitely isn't safe, and it isn't safe for the person vaping or anyone exposed second-hand. However, the evidence so far does suggest the harm is somewhat less than regular tobacco smoke.
The practice is believed to have begun as early as 5000–3000 BC in Mesoamerica and South America. Tobacco was introduced to Eurasia in the late 17th century by European colonists.
Yes. The long-term health implications of vaping are only just beginning to be understood. If you forced me at gunpoint to either vape or smoke, I'd still choose the vape.
Some people have had serious lung damage due to vaping.
If that's a reference to a spate of "popcorn lung" cases of THC vapers, where it's suspected that the juice was laced with vitamin E, you might as well say that Tylenol is deadly because someone once spiked it with cyanide†. Don't put junk in the juice, regulate the ingredients. I'm saying this as a non-vaper and non-smoker, if it makes any difference.
†I know, I know, acetaminophen _is_ deadly is you overdose on it. That's not what I'm talking about here.
Some, but not all [1]: "... the remaining 17% reported using only nicotine-containing vaping products... It is believed that the non-THC/CBD vapers who were diagnosed with EVALI were actually suffering from disparate vaping-associated lung diseases caused by different toxins within the aerosols or different host responses to the inhalants."
Note that many of the toxic chemicals referred to in this paper are normal components of nicotine vape juice, including but not limited to the flavorings, so the comparison to spiking Tylenol with cyanide is not particularly apt.
The comparison to Tylenol spiking is perfectly apt if you consider the whole category of toxic adulterants as the noxious agent. (I know that THC/vit-E wasn't the totality of cases. You can only go so far without drowning in pedantry.) Moreover, it seems to have been a very North American phenomenon. Europe is quite vape-mad and you can count the number of vape-related lung injuries on one hand.
Water is essential for human survival; vaping is not necessary for human survival or health but rather introduces potentially harmful substances into the body, including nicotine, which is addictive, and other chemicals that have adverse health effects.
Serious drowning incidents from drinking water are extremely rare and usually involve underlying medical conditions or very specific circumstances. Conversely, the health risks associated with vaping, including lung injury, addiction, and potential cardiovascular effects, are well-documented and have been observed in a significant portion of users, especially among young people and adolescents.
Vaping has been linked to a significant rise in nicotine addiction among teenagers and young adults, often drawn to flavored vape products. This trend threatens to reverse decades of progress in tobacco control and nicotine addiction. Water consumption does not carry a similar public health risk; it does not lead to widespread addiction or serve as a gateway to the consumption of more harmful substances.
While water is essential and generally safe, there are still regulations in place to ensure its safety (e.g., water quality standards). The call to regulate or ban vapes stems from a similar public health perspective — protect consumers, especially young ones, from products that have been shown to pose significant health risks.
Lastly, equating the act of vaping with drinking water is a false equivalence. The two activities serve different purposes and have vastly different implications for health. While water is necessary for life and generally promotes health, vaping introduces unnecessary risks, particularly to young and vulnerable populations.
That's crazy! I personally think both regular tobacco and vapes should be tightly regulated, with tobacco available only to people born before some fixed year (like New Zealand) and vapes restricted the same way or only to existing smokers. Governments worldwide dropped the ball, allowing an epidemic of vaping, with many of the vapers new nicotine addicts who would never have become smokers had there been no vapes.
Yeah that’s the thing - they massively dropped the ball for several years, the grey market took over, loads of kids started doing it, everybody panicked and now they’re going to be prescription only.
Somewhere in that mess there was a path to making them a good way to really squash that last few percent of people that smoke, at least by moving them onto something a little better even if they never quit, but it’s been missed.
Even then, I’d understand and support making them as hard to get as tobacco, regulated very similarly so when Bob Smoker goes to the servo (gas station) or smoke mart, he can choose the less harmful and less harmful to others path with little effort. Not now though (well later this year), he has to pay for a doctor’s appointment and get an approved therapeutic device prescribed if he wants to vape. Or he can buy cigarettes at the corner.
And the schoolyard mom grapevine apparently has it that vaping is worse than smoking. The clusterfuck is complete, and a great health opportunity was missed.
(For the record I neither vape nor smoke, but I did use the former to quit the latter a little over 10 years ago)
(Edit - oh that thing in New Zealand? The new government scrapped it, I think for reasons of tax revenue. One of the most naked money > lives decisions I’ve seen in recent years)
Totally agree with this. I can see the desire to show off, but I don't understand how anyone can believe this is good marketing strategy. Any initial excitement I get from reading such announcements will be immediately extinguished when I discover I can't use the product yet. The primary impression I receive of the product is "vaporware." By the time it does get released I'll already have forgotten the details of the announcement, lost enthusiasm, and invested my time in a different product. When I'm choosing between AI services, I'll be thinking "no, I can't choose Gemini Pro 1.5 because it's not available yet, and who knows when it will be available or how good it'll be." Then when they make their next announcement, I'll be even less likely to give it any attention.
Matching the 60° FOV of a 27" monitor at close distance, it's equivalent to 1080p or less. It gets even worse at FOVs equivalent to more reasonable distances.
