Reading this article, I just realized that it is something I didn't appreciate enough when I was living in France. In Korea, my home country, you can find a lot of signs for local temples, though. But they do not have identities like French ones. There are some signs too but they are mostly commercial restaurants, local farms, and so on. And they are usually ugly with big letters with saturated colors; red and yellow.
Indeed. I also had this weird feeling while reading through the article. It got hooked up in the beginning. And then at some point, my brain just noticed that it was LLM-generated. I wonder how this article was written. Did the author accidentally find about Voyager 1's tiny memory and its primitive tape technology while reading something else, or did he just ask LLMs to write something interesting that he could publish with a few prompts.
One question. Sonnet for tool use? I am just guessing here that you may have a lot of MCPs to call and for that Sonnet is more reliable. How many MCPs are you running and what kinds?
Nice article. I saw someone depicting the future of web search with AI. The conclusion was not the bright future. Simply put, ads will never go away. Either AI providers will get paid for whitelisting ads, or even worse these AI will directly promote advertised products.
I like the analogy to woodwork and hammer. It fits perfectly to what happens when they do not pay enough attention to the end result. They are not showing the actual product because it is not as shiny as their new agentic hammer.
It matters to me. Claude code is more extensible. They put a lot of efforts to hooks and plugins. Codex may get the job done today. But Claude will evolve faster.
Yeap, it sounds like a big rant with multiple exclamation marks. Having both is a way to go. Recently I purchased a new laptop and thought should I go full Wayland? No way. I started with X11 and then added Wayland. Things break on Linux. You need a stable display server where you can still open a browser, and that is X11. Most of the time, I stay on Wayland until it breaks.
I cannot agree more though I have little experience in open source. I knew that Korean environment for open source software would be touch before coming back from Europe, it seems much easier to target international traction rather than focusing on domestic interest.
Personally, I'd like to know, since you have been active in Korea, if there is any groups that I can attend to.
A friend of mine severely injured her leg, especially knee, and went through a surgery. She said that she had a rehabilitation plan for the next 6 months. Guess what, from Gemini. I just told her just listen to her doctor.
I didn't tell her why LLMs can make mistakes or hallucinate because I thought that she would not appreciate my mansplaining.
Looking forward though, my boring answer would still be education. It is going to take time. But without understanding LLMs, they will not be easily persuaded.
I am in the same boat. It's the unix philosophy. tmux does its job well enough and it is already scriptable. I don't think I was sticking to the same layout for a long time because different projects tend to require different layouts. But that's the fun part of streamlining down your automation. I haven't felt the need to explore other options yet because I haven't felt the limitation of tmux yet.
I am not at all familiar with game development. This article reminds me of Casey Muratori mentioning a font issue in game development environment from a random podcast. On web, you can just fetch a google font whatever. No problem. On a local machine, you tend to look a well-established software like harfbuzz. But then harfbuzz could be rather a big dependency. A game is self-contained and you want your font looks cool and unique to your game, like the Diablo font. So it becomes a design issue. It's an awesome approach to let GPU render fonts. I cannot imagine how many game devs had font issues where they realized that they might have to learn how to render fonts as well not just characters and grass.
Interesting. Can you elaborate more on that? I was wondering the same thing with other providers, if I can scrape data from their platforms. On what grounds do you think it might be okay?
I looked for their APIs not long time ago, which I said non-existent. If you'd like to work with me, I am open to work. If you simply want to talk to me, just drop me an email. I will also checkout your dev tool if I can apply it here.
I also have been brewing a similar idea only in my head. But mine is for the local Korean market. Google maps is simply not reliable. Korean people rely on Naver map or Kakao map, who do not even provide APIs. Users are locked up with these local services. They may launch their own mediocre AIs to help the search, who knows in a few years. But then it's going to be too late. I think they should just open up APIs and let developers explore.
What is worse is a huge number of tourists should also use these local services whose translation is not in place. Soon enough, people will ask their AI assistants which restaurants to try out when they plan a visit to Korea. I can see that there is huge opportunity here. Kudos to Voygr team. Fingers crossed.
I feel often codes from a manger do not go through rigorous code review process, especially if it is a POC. In my case, because of 1 to 10 signal to ratio, it was hard to pin point which parts actually did the jobs done. I am setting things properly this time, me as a maintainer. Thanks for commenting.
This is something even my manager wouldn’t say. I think he is feeling embarrassed from all the evidences. To be fair you didn’t read them. I agree that velocity is everything and as a manager they shouldn’t slow the team down.
As a senior engineer, I am getting extremely tired of reviewing AI slops. Today at work I have decided that I just have to build a POC project from scratch. I spent 2 weeks to review the code, to log the process, and to build toy examples to make my argument clear that some (actually most) parts were not working.
The funny thing is that I know my manager got this “working” within a week with Claude. I had to spend 2 weeks with 4 JIRA tasks, many commits for toy examples, and three reports.
I started seeing this trend nearly a year go in Seoul. I think it is just a sort of fashion trend. I like wired experience when I am alone in a silent place listening to classical or instrumental music. But it’s hard to beat convenience of Bluetooth elsewhere.
My first laptop was xps 13 released in 2016, I think. I am still using it with linux installed. It’s a solid laptop. Good display, good port selection, good keyboard, even trackpad is not bad. It survived my long graduate degrees and survived covid when I was using it full time (mostly ssh though). I swapped the battery two times and battery life is not bad with minimal linux setup. What’s surprising the most to me is that it was just 900 usd.
So it is a startup? I expected it in fact from his reply to my concern. In my opinions, to explore the unknown, I think an institute like Mila, led by Yoshua Bengio, would have been more fitting. But Yann LeCun's career and his reply to my rant[1] speak for himself. I wonder how he is going to make money. Aside all my concerns, I wish him the best.
> You're absolutely right. Only large and profitable companies can afford to do actual research. All the historically impactful industry labs (AT&T Bell Labs, IBM Research, Xerox PARC, MSR, etc) were with companies that didn't have to worry about their survival. They stopped funding ambitious research when they started losing their dominant market position.
Wow. 55 images, all carefully prepared and placed, not a single AI-generated. I love the quality of this post. Not to mention, I learned something new and new perspectives.
Lucky for me I didn't read the sequels. I had my own theories about the purpose of Rama. Some theories are best left as theories in your heads to fuel ideas and imagination.