The models will not be a moat, but the products can be. More specifically "sticky" products / killer apps like ChatGPT, and whatever forthcoming products this acquisition of Jony Ive's company may lead to.
Windsurf acquisition may be explained in part by the same logic of owning a strong and sticky product, as well as a good source of data for training.
Alternatively, this may be Deno’s “Dip”: A tough period of time before continued gains and small breakthroughs that build up over time to a new plateau. Maybe all new creative projects will have this as a part of their journey. I am confident Ryan Dahl is unlikely to give up, and is aware (and working to become more aware) of what is necessary to improve for deno to achieve the vision he has for it.
Dahl doesn't strike me as a business or product person. He's a genius when left to tinker. I get the impression Deno is floundering because of business/VC pressure. I see the original promise of Deno being compromised in an effort to increase users/customers. The project is no longer focused on just making a good JS runtime.
Deno's original positioning was as a second version of NodeJS without the learning cruft cluttering the environment. To that extent I think Dahl and his team was successful.
As is so often the case, once you introduce MBAs/VCs, the focus shifts to ROI and fast. I see Deno Deploy as being part of that attempt.
People still tend to forget that software development tools are not commercially viable. For a long time we have become spoilt for choice with ever more and improving tools.
I read the post from a business lens and an outside observer and this was my hope too. If Deno is buckling down and cutting costs in order to survive a long winter and carry on that seems like the right move for the business and the community at the expense of latency.
This has been the majority of my professional coding experience since the 90's ::shrug::. Having to take ownership of some Java Spring application felt a lot like stepping through an assembly program or some random ROM. Each line of code just returns values and modifies a bunch of other stuff at the same time and you slowly build up a mental model of what it's doing and what it's trying to do!
The way I approach is, I try to design or solve the issue first to the best of my ability, then I let the LLM give me its solution. Compare and take whatever is useful. I know this is not the most effective way, but it's the way I feel like I'm still thinking and not becoming super dependent.
I find that even though I have to spend time understanding what the llm wrote, it's still faster and less energy sapping than if i write the code myself.
Plus I just get the LLM to write tests for it's code and I make sure the coverage is complete.
But it does make me feel uneasy still, like a dirty little cheat.
i've learned a ton from ai suggestions, but right now its like 5-10% of suggestions. If that number was any higher i might not have the motivation to open the black box
... but without all the signs we look for as proof that a given library must not be too bad, since those signs are traces left by developers and by other users of the library, and LLM-generated code doesn't have any other users (I mean, to the extent it's copying existing code, kinda, but it'll be remixed and removed from its original context, so good luck tracking down the source and double-good-luck figuring out if any proof of quality found there is relevant to what you've got)
> One problem I have is that AI is writing code I don't understand completely, yet I'm still responsible for it.
With vibe coding, you are as much a reviewer/editor as an author, and as an editor should never accept generated code that you don't understand. Happily, the same technology that generated the code can explain the code.
No, with vibe coding you are not a reviewer an editor or an author, quoting from the source:
There's a new kind of coding I call "vibe coding", where you fully give in to the vibes, embrace exponentials, and forget that the code even exists.
...
I "Accept All" always, I don't read the diffs anymore. When I get error messages I just copy paste them in with no comment, usually that fixes it.
No, they actually have the files pane on the left, live preview mid-top, terminal errors mid-bottom, and the agent on the right. no looking at code.
Cursor has a "yolo mode" where you don't have to click accept for tooling even for system commands and people whitelist commands like sudo, su, and rf :))) I wish I was kidding.
The vibes are coming from changing to actual product (design, UX, functionality) and not from the code. The code in fact doesn't matter at all. At this point that's only ok for throwaway prototypes (but for those it's quite wonderful), the more the application requires careful maintainable engineering, you need to read every line and leash the LLM. It's a bit of a continuum between the two edges in reality.
> No, with vibe coding you are not a reviewer an editor or an author, quoting from the source:
That was the joke, which Karpathy found "quite amusing" and "not too bad for throwaway weekend projects". In reality, you won't even get that far with vibe coding if you don't understand what's going on.
Considering a library is just a component of the application and not the whole application itself. And one can typically understand the inputs and outputs of a library without having to know the detailed functionally of the code behind the library. I would say yes, it is different.
But this of course can still open up potential vulnerabilities or unexpected behavior. The way I see it, there is some level of trust behind the library that it was built competently
I don't think I've ever seen that actually happen, if you mean reading a meaningful amount of the dep's code.
How many teams importing React have even one member who's ever read even a dozen key paths & files in the React codebase? Think it's more than 10%? I'd bet it's closer to 1% than 10%, of teams on which even one member has done that, ever. Do they check the diff for every update? Shit, IME you're lucky if a single person reads the entire changelog before updating, and certainly nobody's reading the diff. Repeat for everything else. Functionally nobody does this, outside niches where small numbers of low-line-count deps are the norm.
Rails, all those imports from Square you used to have to use to cut the "WT actual F is this?" rate in Android dev to something non-crazy making, et c. Basically nobody even skims this stuff, and I don't think the number of businesses willing to pay for their developers to do that is large.
Like any legacy code, understanding the codebase does not mean reading and memorizing every line in the repo. It's a sense of understanding the rough data flow, then having more intimate understandings of the functions you need at that moment.
Unlike "vibes" there should always be a sense of purpose on why each function is needed and what you plan to do with the output, even if in some cases you treat hat function like a black box in terms of implementation (which IMO should be avoided, but situations may make the implementation above your understanding).
US scientific research funding is largely driven by nepotism and favoritism. Insiders know but don't talk too much about it. They have a few options: a) just quietly stay in the system trying their best to do good work b) join the gravy train through social climbing c) quietly leave and move on with their careers.
This is pretty vague and gestural, I'd love specific examples to support your accusation. I work on NIH funded grants, and while I don't write the grants, I'm reasonably familiar with the process. I disagree with your assessment when it comes to any grant I've been involved with. I've never seen corruption like that. These grant proposals look a lot like private sector bids: here's what we want to do, and how it aligns with your mission, and here's how we plan to do it, and how much we're asking for. The process is competitive, and a committee decides on the outcome. Everything has oversight, and is very procedural. Before working on NIH grants, I worked in the private sector doing large contracts for 13 years, and the downside of the way the NIH does it is not corruption, if anything it's bureaucratic slowness and overcaution. The private sector was much shadier and prone to cronyism, and has nothing to teach the government on that count... believe it or not.
I used to work in academia and was involved in NSF and DOE grants. I’ve been in industry (IC then manager) since then.
My sense is that grant funding was less merit based than industry funding. I’m not saying it’s so corrupt that it should be completely torn down, but there’s just less accountability in academia - you can get a grant, fail to deliver on what you promised, and still get another grant after that and that can be your whole career if you know how to play the academic social game and are good at writing proposals.
The whole point of many of these grants is to invest in research on the leading edge where, by definition, nobody knows whether what they're trying to do is going to work, because it's actually new.
Of course that can be gamed, and of course we need good faith oversight, but if none of the research projects we're funding were to ever fail, that would be evidence that we're being massively too conservative in the avenues for new discoveries that we're investing in exploring.
If that's true, and you've offered no evidence to support that it is, canceling funding for anything tangentially associated with "DEI" doesn't seem like it actually solves any of those problems.
Please name a specific government grant that was given to a specific researcher based nepotism and favoritism.
The last time I checked when I worked at a Stanford biomedical university department that was substantially NIH-funded, there were 2 full time employee grant writers who had to supply the government grant process with a laundry list of specific data with each carefully-worded proposal because they were regularly competing with other universities to win a specific grant.
Some Stanford guy caused the NIH to deprioritize the infectious etiology theory for Alzheimer’s for decades. It’s not clear why his research was considered better than the others’, but it did receive a lot of attention and became the driving force. Millions have paid the price. His voice was not the only one, it was just the favored one.
Insiders have little say. NSF is probably the most merit based system in all the US government. Literally any other program (defense?) is less merit based.
Also, if nepotism and favoritism are the criteria for removal, let's start with the Executive branch.
Even taking this as true, and also taking it as true that we need to build a new research funding system from a clean slate to fix it - that can still be done in parallel while leaving the existing system in place! So as usual the effort of trying to discern some higher purposes is unwarranted, and the goal of these people is to just straight up destroy our country.
I had the advantage of having a mentor early in my career hold my code to the highest standards with regards to using the semantics of HTML. We were working on a big redesign of a large website for BigCorp, but we were still a small team on a deadline. She would not accept any excuses, and she guided me to think through HTML element selection until we found what we considered the best choice. I was taking the bus to work most days and used the time to read on the bus and at home the thick book "CSS: The Definitive Guide" by Eric Meyer cover to cover.
As you mention, working with HTML, and even more so CSS, can be a source of frustration. The UX of actually working with them is tedious work. However I can write today that all these years later, the high standards that my mentor held me to (and the project required) helped me to master HTML and CSS in a way that made doing that work less tedious, and easier over time. I was being paid to be educated by an expert - I considered it a great opportunity, and believe I have been proven correct.
In the post Vasilis writes that they told their students the assignment "doesn’t have to be semantic and shit". I consider this a missed opportunity to hold them to higher standards and help the students build a strength that can help them for years to come. More broadly, I believe this to be an example of current generations being limited by their mentors and educators lowering standards, and potentially robbing them of opportunities. Impedimentum Via Est.
I'd argue that the semantics of <bigcorp-career> vs <into-webdev-class> are different enough that it's a bit of an antipattern to apply the same style to each.
RFK Jr and Nicole Shanahan are running, and already on the ballot in 29 states. Though you won't see them interviewed on CNN or Fox News, many polls show them winning against Trump.
One of the reasons that RFK Jr started his campaign last year was a poll by Zogby which "surveyed over 26,000 likely voters across all 50 states, indicated that Kennedy could potentially outperform both President Joe Biden and former President Donald Trump in head-to-head matchups." I believe the poll showed the majority of independent votes going to RFK Jr, and a significant portion of D and R votes going to him as well.
Given recent events, perhaps Trump is now much more popular, but previously polls showed RFK Jr beating Trump.
It’s more conceivable than people are willing to believe that RFK could get more votes than trump in this election. For example, trump might die before election day, in which RFK might have a chance.
> perhaps Trump is now much more popular, but previously polls showed RFK Jr beating Trump
Trump isn’t more popular. RFK was never beating anybody. If he played his card better he might have put himself into the running for Trump’s VP.
The bigger problem is there is a strong political incentive to look at who doesn’t vote and transfer their chips to those who do. (Appealing to non-voters doesn’t work. Most non-voters aren’t principled, they’re just lazy or indecisive, irrespective of what they say.)
The models will not be a moat, but the products can be. More specifically "sticky" products / killer apps like ChatGPT, and whatever forthcoming products this acquisition of Jony Ive's company may lead to.
Windsurf acquisition may be explained in part by the same logic of owning a strong and sticky product, as well as a good source of data for training.