Something about people successful with computers makes them quick to claim something is easy based on the number of steps needed, without regard to the ease of remembering all the arbitrary or sometimes contra-pattern steps required.
In my defense, I remember it because I expect Option key to modify the original character and Shift key to make it bigger, so remembering that Option plus Shift makes hyphen into a bigger alternate version of it, i.e. the em dash, is not difficult. I acknowledge that not everyone would see it this way.
First of all, thank you for making this. I used it in my recent job search and it was fantastic.
Second, if I may make a request, could you please follow SemVer? I tried rendering my resume again last week, only 3 or 4 months after having made it originally with RenderCV version 2 point something I cannot recall, and it would not work. The design schema and perhaps also the CLI options have changes so much that I expect I would need to spend 2 to 4 hours getting it to work again, and there is no guarantee that it would not break again in another month. I would have appreciated if the versioning scheme followed SemVer, so I would know that any v2 engine would work and v3 engine would not.
I also would appreciate it if you could write detailed migration docs between versions and/or recommendations in error messages. The reason I think migrating my CV would take so long is that I have to go by trial and error, searching for similar-sounding parameter names and replacing them one-by-one. I gave up after an hour of this as I was nowhere near done.
Third, is markdown render supposed to miss information or is it a bug? Some sections of the resume would not end up in the markdown version, only showing section title and nothing else. If this is not expected behaviour, please let me know.
Again, thank you for making this. I look forward to using it again in the future.
Thank you so much for the kind words, and I'm sorry for the trouble the changes caused you.
To give some context: I hadn't worked on RenderCV for about six months, and when I came back, I had grown technically and my design taste had changed. I decided to do a significant overhaul rather than preserve backwards compatibility. It felt necessary to maintain my enthusiasm for this project long-term.
You're right about semantic versioning. I used a two-number scheme (MAJOR.MINOR) back in 2023 when I didn't know much about releasing software. By the time I understood the benefits of MAJOR.MINOR.PATCH, I didn't want to change it mid-stream. I'm planning to switch after v3. It will give me more flexibility to tag updates appropriately.
I'm also adding migration documentation to my list. You're right that I should have done this. Going forward, there will be clear migration guides.
Regarding the Markdown issue where sections show only titles, that sounds like a bug. If you could open an issue with details, I'd appreciate it.
The author discovered AI coding 2 weeks ago and completely went to town on the entire project. If you use any version before this (2.2) it will probably work fine.
Thank you for writing this post. Your writing is insightful and thought-provoking. I would love to follow your blog to read your future posts as well, but I could not find an RSS feed or an email newsletter option. Is there any chance that you would add RSS to your blog in the future?
I can see the logic. Layoffs are always terrible. But if I am getting laid off anyway, I would prefer to know about it before I spend a whole bunch of money during holidays.
I have never had a heat pump, so I wasn't aware of this shortcoming. Could you please explain a bit more how different it is with heat pump compared to furnace?
The heat pump will always produce air that is warmer than the temp in the house, but as the temp outside drops the temp of the air coming out of the vents also drops. So on a very cold day when the house temp is say 70F, the system might only be putting out air that's 75-80F. The air coming out of the vents doesn't really _feel_ warm and it may take an hour or two to raise the temperature in the house when you wake up or get home in the evening.
In my experience at least with relatively modern heat pumps (roughly 2000 and newer) it doesn't matter that much when outside temps are above freezing. But it quickly starts to become noticeable as temps drop into the 20s.
I see. Thanks for the explanation. So the system is slow to come up to the set temperature. Is it good at keeping the temperature though? After the house temp gets to 70, does it consistently stay at 70, or are there shortcomings in this aspect too?
Also Canadian here. Have been closely following US-Canada trade for years. What you are saying is news to me. I am leaning towards believing you are simply mistaken, but if you have specific evidence about these two assertions, I would love to study them:
> tariffs on a ton of stuff
> we are rebuked because we have tariffs on a bunch of their goods
I can guess why one might think some of this (tariffs on Chinese EVs directly led to agricultural counter-tariffs from them and dairy trade barriers have always been a source of frustration), but in general, Canada's tariffs and barriers are by all indications in line with peer countries (US, UK, EU, Australia, etc.) and not particularly noteworthy. If you have concrete evidence to the contrary (not just that some trade barriers exist between Canada and its trading partners, but that they are out of the ordinary and much higher than other countries'; and that the world, in particular EU does not want closer ties with us because of them), I would love to study your sources and update my understanding.
A 'digital twin' is a model that evolves over the lifetime of a physical asset. So if the state of the real device changes, so does the model (data assimilation). Likewise, one can use the model for control/planning/"what-if" scenarios. This is the bi-directional information flow that's being mentioned.
So what I'm showing on the website is just the model part. I'm not a fan of exposing my hardware in a public demo (the digital twin part), but the idea is that this model evolves with the roaster during the roast (data assimilation) and can help the operator guide the roaster to a desired end goal (e.g. medium roast along some profile or with minimal energy usage).
> A digital twin is a digital model of an intended or actual real-world physical product, system, or process (a physical twin) that serves as a digital counterpart of it for purposes such as simulation, integration, testing, monitoring, and maintenance.
> A digital twin is "a set of adaptive models that emulate the behaviour of a physical system in a virtual system getting real time data to update itself along its life cycle. The digital twin replicates the physical system to predict failures and opportunities for changing, to prescribe real time actions for optimizing and/or mitigating unexpected events observing and evaluating the operating profile system."
The only effect of applying to jobs on jobs.now is that someone who is already working in the US on H1-B, and was gonna get green card if you hadn't applied, would instead stay on H1-B.
Another thing is that the company must respond to your application within 30 days. Unlike Indeed or Linkedin applications which simply disappear into a void.
H1-B visas eventually expire. I think it’s technically a guest worker visa, so the expectation is that the visa holder will go back to their home country after working in the US a few years.
H1-B is an immigrant intent (actually dual intent) visa. I know a lot of Canadians that started working in the US on a TN visa, which doesn't let you ever become a permanent resident. They had to do all the paperwork (and win the lottery) to convert to H1-B status so they could begin the process of getting permanent residency.
To be clear, H1B is not an immigrant intent visa. It’s a non-immigrant visa. “Dual intent” is like Schrödinger's intent. It allows someone who is on a non-immigrant visa to avoid the presumption of immigrant intent that would otherwise apply—rendering them deportable—when they apply for a green card.
The statutory protection for H1Bs is thin. In 1990, Congress excluded H1B from the requirement applicable to other non-immigrants that they retain a foreign residence, and from the rebuttal presumption that someone who applies for a green card has immigrant intent. That’s it. The common operation of H1B as being an immigrant-intent visa is mostly a matter of administrative grace.
The system is byzantine because it is a hack around public opinion. Since 1965, support for more immigration peaked around 35%: https://nypost.com/2022/08/08/more-americans-want-fewer-immi..., with everyone else wanting it to either stay the same or be reduced.
So Congress can take advantage of the fact that a lot of people are okay with the status quo, but can’t affirmatively enact legislation that would be seen as creating a new pathway for permanent immigration.
So the current system was built on a series of small measures that could fly under the radar. Eliminating the rebuttable presumption of immigrant intent in subsection blah blah blah doesn’t sound like it’s going to create a new pathway for permanent immigrants.
It's six years in total, excluding time spent outside the US. The status is valid for at most three calendar years, but it can be shorter for fixed-term positions. And you can renew it as long as your total time in the US is less than 6 years.
I was on H-1B at a university where researcher appointments were nominally from July of year N to June of year N+2. But if you didn't start in July, your second appointment might be only 1 year, for some bureaucratic reasons. And you had to renew the H-1B for each appointment. I had five H-1Bs in total over ~7 years.
The author would still be wrong in the tool-calling scenario. There is already perfect (or at least superhuman) chess engines. There is no perfect "coding engine". LLM's + tools being able to reliably work on large codebases would be a new thing.
Correct - as long as the tools the LLM uses are non-ML-based algorithms existing today, and it operates on a large code base with no programmers in the loop, I would be wrong. If the LLM uses a chess engine, then it does nothing on top of the engine; similarly if an LLM will use another system adding no value on top, I would not be wrong. If the LLM uses something based on a novel ML approach, I would not be wrong - it would be my "ML breakthrough" scenario. If the LLM uses classical algorithms or an ML algo known today and adds value on top of them and operates autonomously on a large code base - no programmer needed on the team - then I am wrong
Not on Mac:
hyphen/dash: -
En-dash: ⌥-
Em-dash: ⇧⌥-
reply