keeda

34 ms·

361.
by keeda 6mo ago
I never shipped an app in VB6 (only dabbled extensively) but I did work professionally with a huge range of UI technologies, including vanilla HTML/JS, ColdFusion, Java Applets, AWT/Swing, C++/QT, Flex/ActionScript, Obje
362.
Token spend breaks budgets – what next?(newsletter.pragmaticengineer.com)
3 points by keeda 6mo ago|0 comments
363.
by keeda 6mo ago
Relevant Pragmatic Engineer newsletter with many more cases along these lines, along with how some people are handling them: https://newsletter.pragmaticengineer.com/p/the-pulse-token-s... Tokenmaxxing seems more and m
364.
by keeda 6mo ago
True, but all the hyperscalers and neoclouds have been severely capacity crunched for multiple quarters and have a backlog of a trillion+ dollars. So even if Anthropic wants capacity it's going to be a) hard to come by (like Dario said
365.
by keeda 6mo ago
> That he’s a liability to OpenAI, which is slowly coming around to the realization that it would be worth more without him. I'm curious what you're basing this on. Are you aware of any grumblings on the inside? From the outs
366.
by keeda 6mo ago
Yeah, this is what I surmise too, but I do have doubts about how lucrative this model will be after the eventual transition to agents. Basically my thinking is this: 1. Google has monopolized the current AdTech stack from all sides. Not o
367.
by keeda 6mo ago
Makes sense, and that was the surprising thing about WoW using TCP. I wonder if Blizzard chose to put in all that extra effort to make TCP work because they encountered enough crappy home routers out there that mangled any non-TCP traffic..
368.
by keeda 6mo ago
Agreed, the techniques in general (caching, backpressure, exponential backoffs, etc.) are well-known, but a couple of things: 1) The general cause of issues in these cases is that certain assumptions no longer hold, and above a certain leve
369.
by keeda 6mo ago
I wrote this in response to the below comment, which is now edited and unfortunately dead, so posting here: I understand, that wasn't a comment on your efforts back then, just that it is a solved problem today. But that does not mean o
370.
by keeda 6mo ago
The Forward Deployed Engineer role, as envisioned by Palantir, is qualitatively different from the other roles like partner engineers or solution architects, and I think it will be critical for enterprise AI adoption. I found this blog from
371.
by keeda 6mo ago
This is the talk by Carlini, only half-way through it but matches what you described i.e. run the prompt on each file: https://www.youtube.com/watch?v=1sd26pWhfmg
372.
by keeda 6mo ago
It really, really depends on what you mean. Specifically, it depends on the application and its various compute, I/O and access patterns. Scaling ecommerce and games is well-known by now (e.g. Amazon and Blizzard have been dealing wi
373.
by keeda 6mo ago
To be fair, the initiators of the "Open Source" movement also co-opted a term that previously had a much more flexible meaning (and had been around for more than a decade at that point.) Just writing a document attributing specifi
374.
by keeda 6mo ago
I'm curious, do you recall if they gave any technical details about how they thought about AGI? Like, was it based on neural networks or something else, like symbolic AI? Asking because, reading the tea leaves from the outside, until C
375.
by keeda 6mo ago
If you look into how she ran the agency, there were a lot of parallels with how the current US adminstration is being run. (Ask any AI for an overview and see the parallels pop up.) Regardless of your political leaning, I think generally we
376.
by keeda 6mo ago
If you look at the number of countries that require licensing to practice, there are literally a handful. It doesn't make sense to dismiss the vast, vast majority of engineers in the world as "not really engineers."
377.
by keeda 6mo ago
I will preface by saying that someone with Lina Khan is sorely needed; Big Tech and other monopolies have gotten way too Big and seriously need to be reined in. That said, from all the informed takes I've seen, Lina Khan was seriously.
378.
by keeda 6mo ago
I'd agree that applies to people, or more accurately specific projects, but not the discipline of software engineering as a whole. Even most of the projects I personally have worked on simply did not need "engineering" as suc
379.
by keeda 6mo ago
>* ... legally enforced ...* Other than that part (most countries in the world do not have regulations or licensing requirements for most engineering disciplines) I would agree. But I would also point out the set of software projects tha
380.
by keeda 6mo ago
Most countries do not need accreditation for engineers.
381.
by keeda 6mo ago
> ... nobody in the field of "Software Engineering" is actually an Engineer if we go by a strict definitions. This is a pet peeve of mine, so while I understand what you mean, I will challenge you to come up with a strict def
382.
by keeda 6mo ago
Missed opportunity to call it AIzheimer? ;-)
383.
by keeda 6mo ago
That's fair, largely because an LLM is a lot more capable at overcoming restrictions, by hook or by crook as TFA shows. However, most systems today are not even resilient against what humans can do, so starting there would go a long wa
384.
by keeda 6mo ago
Agreed, but the point is, if your system is resilient against an eager intern who has not had the necessary guidance, or an actively hostile disgruntled employee, that inherently restricts the harm an LLM can do. I'm not making the cas
385.
by keeda 6mo ago
Actually I think the opposite advice is true. Do anthropomorphize the language model, because it can do anything a human -- say an eager intern or a disgruntled employee -- could do. That will help you put the appropriate safeguards in pl
386.
by keeda 6mo ago
I think that may have been a result of the political divide of that era. The USSR did export some machinery and arms, but those were traded largely within other Communist countries and "third world" countries.
387.
by keeda 6mo ago
> Nobody has found a better process. Not even close. Maybe, but as they say, "Past Performance Is Not Indicative of Future Results." The point was, this process may work up to a point, then it may work more against the intere
388.
by keeda 6mo ago
I think it's more accurate to say it is a process that has resulted in our high standard of living faster than other processes ... so far. There is no guarantee it will keep working for the majority of us going forward; as is becoming
389.
by keeda 6mo ago
Actually I have the opposite take. This is largely a play to procure compute capacity (and I suspect, distribution via Google Cloud), and I think Dario wildly underestimated the amount of demand they would see. I always wondered why Anthrop
390.
by keeda 6mo ago
My personal theory, represented in the points in TFA but not quite pinpointed, is that it is due to smartphones and the overall media environment (not just social media). Specifically: Smartphones enable unprecedented levels of reach as w

More ›