"Jobs are going to be lost unless there's somehow a demand for more applications."
That's why I'm not worried. There is already SO MUCH more demand for code than we're able to keep up with. Show me a company that doesn't have a backlog a mile long where most of the internal conversations are about how to prioritize what to build next.
I think LLM assistance makes programmers significantly more productive, which makes us MORE valuable because we can deliver more business value in the same amount of time.
Companies that would never have considered building custom software because they'd need a team of 6 working for 12 months may now hire developers if they only need 2 working for 3 months to get something useful.
I built an Excel Add-In that allows my girlfriend to quickly filter 7000 paper titles and abstracts for a review paper that she is writing [1]. It uses Gemma 2 2b which is a wonderful little model that can run on her laptop CPU. It works surprisingly well for this kind of binary classification task.
The nice thing is that she can copy/paste the titles and abstracts in to two columns and write e.g. "=PROMPT(A1:B1, "If the paper studies diabetic neuropathy and stroke, return 'Include', otherwise return 'Exclude'")" and then drag down the formula across 7000 rows to bulk process the data on her own because it's just Excel. There is a gif on the readme on the Github repo that shows it.
I hate to be flippant, but it's just abundantly clear that Pahlka is soon going to be the newest recipient of a "fell for it again" award. I'm still an Elon fan at heart, but it's absolutely impossible to believe he's interested in the nuts-and-bolts view of government efficiency she's describing. He's repeatedly said that the problem is the government spending too much money, which he'll fix by finding employees, departments, and regulations to get rid of.
The problem with this argument is that you are putting TikTok on a pedestal.
The US (and every country on the planet) has rules and regulations around who can do business in their country and who their citizens may do business with.
If you want to argue that the U.S. shouldn't be able to prohibit its citizens from doing business with TikTok you should spend some time generalizing that argument and figuring out a good reason we shouldn't be able to prohibit Americans from selling weapons to Russia, or allowing Russian companies to set up manufacturing facilities in the United States to build weapons to send back to Russia. (or any other scenario you want to make up)
"Free speech" is not a good argument here. TikTok isn't a "free speech" platform. It's just a random company selling products and services in the United States.
1. Always use `git switch` instead of `git checkout`
2. Avoid `reset --hard` at all costs. So for the "accidentally committed something to master that should have been on a brand new branch" issue, I would do this instead:
# create a new branch from the current state of master
git branch some-new-branch-name
# switch to the previous commit
git switch -d HEAD~
# overwrite master branch to that commit instead
git switch -C master
# switch to the work branch you created
git switch some-new-branch-name
# your commit lives in this branch now :)
3. I'd apply the same to the `cherry-pick` version of "accidentally committed to the wrong branch":
git switch name-of-the-correct-branch
# grab the last commit to master
git cherry-pick master
# delete it from master
git switch -d master~
git switch -C master
4. And also to the "git-approved" way for "Fuck this noise, I give up.":
# get the lastest state of origin
git fetch origin
# reset tracked files
git restore -WS .
# delete untracked files and directories
git clean -d --force
# reset master to remote version
git switch -d origin/master
git switch -C master
# repeat for each borked branch
ORMs are a bad idea, even when using OO langs: they make the simple queries slightly simpler (`Users.getById(id: Long)`), they do not help you for hard queries (ORM-using codebases of size usually have hard-SQL-queries "in strings").
Most users of FP langs know this and hence will not even try to implement ORMs.
Look into jOOQ, LINQ-method-syntax (or whatever it is called, without the funny SQLish syntactic sugar), SQLDelight or sqlx for non-ORM options that improve embedding SQL in general purpose langs.
Is it censored, or do most people just not talk about it on Facebook?
It's interesting how incredibly supportive of human rights that a platform in bed with the CCP became, no? Do you think that China's human rights bugaboos are often discussed on their internal social networks?
I ran a YouTube channel seriously for a couple of years. Chasing the whims of the algorithm contributed to burnout and now I only release one or two videos a year.
My reason for starting a channel was to have a discussion around some of the projects I was working on that I found interesting. This never materialized. I never had a ton of engagement, but if I did most questions asked by viewers that are answered in the video.
I monetized my videos and started making beer-money amounts of revenue. I put more and more time into them but never gained traction. I had a few "evergreen" videos that would make maybe a thousand bucks a year, the rest of the videos hardly got any views at all.
Eventually, I found myself making videos to feed the algorithm not because it was a project that I wanted to do. This is where I had to stop. I realized that I just don't like editing videos.
My channel makes no revenue now because YouTube requires an upload schedule and shorts. Mr Beast's job sounds awful to me. The videos I make now only get dozens to hundreds of views when they're published. I mostly share them with friends and in online communities where the info contained might be helpful.
Without commenting on this subthread (I don't have an opinion), you or anyone else with this concern should look into sticky modifiers (modifiers that apply to the next key press without being held). They were a game changer for me personally as far as managing RSI, as I had a bad habit of tilting my wrist to eg type a capital T.
Again, DEI does not mean "hire less qualified people from group X", it mean "don't bias your hiring of less or equally competent candidates in favor of men". It is not "sexist/racist in the other direction" - though I'm sure some people can make it such - however, let's be clear here, if it _were_ racist/sexist in the other direction, and you want it to be removed, you're saying "both paths are sexist/racist and I just prefer the old sexist/racist path that gave an advantage to people who are straight, cis, white, and/or men". It's really not the most compelling argument to say "I just wish we got to benefit from the biases we used to have".
The problem is lots of people are use to a default world in which being a man, or being white, etc means that if all other things are equal they will be selected. They see anything other than choose the "equally performing" white dude by default as reverse racism.
Couple this with the well documented history of women and minorities being judged more critically for the same work, and you get the "non-racist" world where worse performing candidates are hired because they're "equal" and a better culture fit.
You can see the same thing retroactively: a woman or minority figure who subsequently performs below the level expected is blamed on "DEI", yet no one questions the hiring of a sub-performing non-minority. Weird right? lackluster minority employees only ever get their jobs because of DEI, lackluster non-minorities are just a thing that happens some times.
You're right, it would be great to not consider gender, race, etc - and certainly when I've interviewed people it has never figured into the process, but companies need to have some kind of mechanism to compensate for all the employees that do (you can happily find countless examples on HN of people openly admitting they consider women inferior devs, they dismiss certain races and castes, lgbt candidates, or people with the wrong religion. DEI programs exist so that those bigots can't pollute the entire recruitment process. Those are the bigots who complain most, yet seem completely unaware that they are the entire reason DEI programs are necessary. The moment you say "I don't think group X can do the job as group Y", you've immediately said there needs to be an external mechanism to mitigate the advantage you are giving to group Y.
That's what DEI is: a hamfisted, but sadly necessary, unless you have a better idea, mechanism to stop people from biasing the recruitment process.
Seriously.
If you think DEI is bad, I want you to think about every "bad" coworker you've had and see if there's any bias in what you blame for them being hired. If you look at the news today, everytime anything goes wrong, and there is a minority involved anywhere, even if they were objectively not at fault, the bigot circle jerk does nothing but blame DEI.
Lots of people want to go back to the "good old days", because they were immensely privileged in those days, you could get and stay married easier - because women could not have bank accounts, and were functionally owned by their family or husband. Remember a women could not divorce her husband for beating her, nor for raping her, because by definition spousal rape was impossible. So yeah, I understand why lots of men want to go back in time, life was objectively easier for them. Those men just don't give a shit about it being harder for literally every one else, which is why I don't give a shit about their opinion.
If you think DEI is the reason you can't get a job, maybe the problem is just that you're broadcasting "I'm a bigot" and no one wants to work with someone who is saying "I have no respect for my coworkers unless they're the same race, gender, religion, etc as me and will not treat them as people". Or of course, maybe it's just dunning-kreuger: you're too low skill to understand you're lower skill than the people being selected over you.
A lot of web frameworks etc do most of the instrumentation for you these days. For instance using opentelemetry-js and self hosting something like https://signoz.io should take less than an hour to get spun up and you get a ton of data without writing any custom code.
It is as complicated as you want or need it to be. You can avoid any magic and stick to a subset that is easy to reason about and brings the most value in your context.
For our team, it is very simple:
* we use a library send traces and traces only[0]. They bring the most value for observing applications and can contain all the data the other types can contain. Basically hash-maps vs strings and floats.
* we use manual instrumentation as opposed to automatic - we are deliberate in what we observe and have great understand of what emits the spans. We have naming conventions that match our code organization.
* we use two different backends - an affordable 3rd party service and an all-on-one Jaeger install (just run 1 executable or docker container) that doesn't save the spans on disk for local development. The second is mostly for piece of mind of team members that they are not going to flood the third party service.
[0] We have a previous setup to monitor infrastructure and in our case we don't see a lot of value of ingesting all the infrastructure logs and metrics. I think it is early days for OTEL metrics and logs, but the vendors don't tell you this.
This take is cynical to the point of wilfull ignorance. My spouse works in DEI and I guarantee you her and her coworkers are sincere and trying to instill better, less biased hiring practices and to make everyone feel welcome and part of the team. Not everyone is going to be the same but that's like anything else. Being 100% dismissive is as much of a mistake as being 100% unquestioningly accepting.
My favorite work in this vein is still the guy who did the trigonometry to convert a year+ worth of telemetry from a light level sensor in his back yard to paint a picture of the tree canopy. Each day is a scan line of light and dark patches at a different solar inclination. So everything above the sensor eventually got “painted”.
My Christmas Eve tradition with my wife for the last 14 years has been to eat as much Chinese food as we possibly can at our favorite Chinese place. Then we drive around town looking at Christmas lights while listening to classic Christmas music. Neither of us are religious, but we were both raised Christian.
I came into this thread thinking that I don't really have any great Christmas memories. My family was poor growing up, my parents shouldn't have been together for as long as they tried to make it work, and my father had a lot of issues with drugs and alcohol. But now that I think about it, these Christmas Eve memories with my wife are my favorite.
Seems like the tag system is flat, which is a big limitation on the organization capability.
For example, I noticed that in the demo access app, there's a note about cooking, and it has 4 tags:
- `baking`
- `cupcakes`
- `oven cooking`
- `recipe`
This would get out of hand quickly.
There should be a hierarchy of tags (categories):
`cupcakes` in `baking` in `oven cooking` in `recipe`
The only tag needed in this case for the note would be `cupcakes`
Ro Khanna said something similar a week or so ago. You all get that this would basically be the biggest corporate giveaway in the history of American politics, right? The reason you're getting health care denials is that providers are trying to rip your face off. I'm pleading with people on this site to go look at where the dollars in health care spending actually go. CMS, the government organization that manages Medicare (and somehow got vilified a few weeks ago when Blue Cross tried to adopt Medicare's anesthesia billing rules), tracks this stuff in a set of documents called the National Health Expenditures. They're spreadsheets, they're easy to go read, please, for the love of all that is holy, go look at them before designing policy.
haha this is gonna be an interesting one for sure.
it's crazy how us humans yearn for magic pills for the highs and the lows just to get ourselves in a healthy baseline
and as bryan johnson's experiments look like, it all comes down to our consistency in eating healthy, exercising regularly, and resting enough - those three alone will net you better health than most people - and it's hard to get into habit of doing those since we all got distractions around us.
In that vein, perhaps the delta between o3 @ 87.5% and Human @ 85% represents a deficit in the ability of text to communicate human reasoning.
In other words, it's possible humans can reason better than o3, but cannot articulate that reasoning as well through text - only in our heads, or through some alternative medium.
Surely not all that blood is on the hands of the CEO, or even UHC. But everyone has either been denied claims or knows someone that has, through some of the most disgusting treatment witnessable from corporate america. It'd be one thing if you were denied a claim for.... any variety of reasons that follow from the inherent difficulty of managing health insurance as many countries can attest to. But overriding doctor concerns to withhold funding for necessary treatment and withholding as much as possible while raking in profits leaves people with very bitter tastes in their mouths.
Mine included: I'm paying for a medical procedure for a friend next week because her insurance won't cover anesthesia and she can't afford the out of pocket. I had multiple relatives die much sooner than they had to because they were denied funding for critical medication. I've had multiple friends talk about "rationing insulin" (thankfully, both times referring to the past). And naturally healthcare interests—UHC included—lobby both parties heavily, so there's a lot of incentive to continue ignoring the elephant in the room.
ClickHouse is awesome, but there's a newer OLAP database in town: Apache Pinot, and it is significantly better: https://pinot.apache.org/
Here's why it is better:
1. User-facing analytics vs. business analytics. Pinot was designed for user-facing analytics (meaning analytics result is used by end-user (for example, "what is the expected delivery time for this restaurant?"). The demands are much higher, including latency, freshness, concurrency and uptime.
2. Better architecture. To scale out ClickHouse uses sharding. Which means if you want to add a node you have to bring down the database, re-partition the database and reload the data, then bring it back up. Expect downtime of 1 or 2 days at least. Pinot on the other hand uses segments, which is smaller (but self-contained) pieces of data, and there are lots of segments on each node. When you add a node, Pinot just moves around segments, no downtime needed. Furthermore, for high availability ClickHouse uses replicas. Each shard needs 1 or 2 replicas for HA. Pinot does not have shards vs replica nodes. Instead each segment is replicated to 2 to 3 nodes. This is better for hardware utilization.
3. Pre-aggregation. OLAP cubes became popular in the 1990s. They pre-aggregate data to make queries significantly faster, but the downside is high storage cost. ClickHouse doesn't have the equivalent of OLAP cubes at all. Pinot has something better than OLAP cubes: Star trees. Like cubes, star trees pre-aggregate data along multiple dimensions, but don't need as much storage.
I have a friend who's a pastor. As a basic rule of a good sermon he always quoted the old addage: "Tell them what you’re going to tell them, tell them and tell them what you’ve told them".
IMO Bluesky (really atproto) is decentralized along the axes that matter, while Mastodon is decentralized along the axes that don't. It's sort of a figure-ground inversion in thinking about a social media protocol so I think a lot of the criticism is coming from people who haven't taken any time to understand it.
It's an open protocol, but more like the www than email. You can port your identity to another host at any time. You can self host if you want. You have complete control over how your chosen feed algorithm aggregates posts from the firehose. If you want to make a competing service to Bluesky you can even do that, and it will still interoperate with all the Bluesky users.
Whereas with Mastodon, when you pick an instance you're essentially picking which benevolent dictator you want mediating your experience. Mastodon is decentralized in the sense that it breaks the platform up into smaller fiefdoms; Bluesky is decentralized in the sense that you retain control over your own experience.
Your edit demonstrates exactly what I mean by bad faith: you're intentionally itching for a fight in all your comments, you're not focused on conveying your points in a way that invites level-headed discourse.
Are things ok with you in general? Could there be some non-political reason why you're resorting to inflammatory techniques to get people to talk with you?
That's why I'm not worried. There is already SO MUCH more demand for code than we're able to keep up with. Show me a company that doesn't have a backlog a mile long where most of the internal conversations are about how to prioritize what to build next.
I think LLM assistance makes programmers significantly more productive, which makes us MORE valuable because we can deliver more business value in the same amount of time.
Companies that would never have considered building custom software because they'd need a team of 6 working for 12 months may now hire developers if they only need 2 working for 3 months to get something useful.