ctvo
This is the typical "ugh that's just rsync" when Dropbox was announced, but for a new LLM age.
_If it's so easy, do it_. Don't post about it. Do it. Fix these bugs at scale. Charge people money for it once you've built up a reputation. Get your own bug bounties.
Yes, exponential backoffs alone are insufficient. Yes, adding jitter helps randomize the calls across a fleet and should be the default with exponential backoffs. Yes, both of these may be sufficient for most systems. Yes, you can dive more into circuit breakers and adaptive retries to limit thundering herd.
https://aws.amazon.com/blogs/architecture/exponential-backof...
Frontier models being commoditize is inevitable. OpenAI thinks they're still competing on technology, and not user experience and market reputation otherwise they'd understand the continuous negative PR generated by Altman's chaos is going to cost them everything.
Repeat after me:
_Other people can experience things you do not experience and it is still valid, and not a delusion_. They are not sheeple who fell for marketing.
A private business can 100% refuse service to you. Examples with regards to "delegation":
- If you come in using a form of non-cash payment that doesn't belong to you.
- If you're purchasing a car, and are filling out paperwork under someone else's name. FYI, you can buy cars on Amazon.com.
- If you attempt to pick-up a pre-order or an item earmarked for someone else.
...
Of course some businesses are more or less restrictive base on fraud chance, yada yada, but you get the idea. You're not being oppressed. Go shop elsewhere.
What is the regular API? How do you express all the integrations needed in this API? Who provides the integrations? Answering these questions lead you back to something like an MCP, which is an API contract that can be as generic or as specific as needed. Wasting context window to understand and re-implement each integration is why MCPs exist.
All the security issues are orthogonal, and occur regardless if invoking this API occurs via code or natural language.
_Telling the browser how you want the DOM manipulated_ isn't the expensive part. You can do this just fine with Javascript. The browser _actually redrawing after applying the DOM changes_ is the expensive part and won't be any cheaper if the signal originated from WASM.
Front-end engineers have no issues adopting new frameworks. See the common complaints about the speed front-end stacks change vs. say Spring MVC or Rails.
A more interesting examination is what is the impact of agentic AI tools being able to write better, more idiomatic code in React vs. Svelte because there's more of it. The human side is less of a barrier here.
How does this change the point? They would still like the best candidate out of that pool, not any warm body, since they have limited positions.
What is your approach to hiring and evaluating talent knowing the large number of applicants and how easy it is to _talk about software development_ vs. _actually developing software_, and how expensive and difficult it is to deal with a bad hire, even in America.
Has Gamer Nexus produced shoddy work in the past? The comparison feels lazy.
I’m more than happy to read a short sellers investigative piece, for example. You don’t need to be journalists to produce good work backed by facts.
Outside of that use case, I enjoy using Kagi and recommend it to most people.
The safest space continues to be distributed systems and systems programming in general IMO. You'll still find work at hyperscalers. You'll still find work in the new spaces. Until AI can operate these systems, there'll be a spot for us bit janitors for a little while longer.
My eyes rolled so far back I hurt myself.
Please provide some examples of where the EU has been able to do a fraction of what you listed to large, US based firms in the past.
Looking at the future, if you want a trade war and an excuse for the new US administration to completely neglect NATO obligations this is a great start.
That's partially the cost, but the other cost is building this into existing tool chains and deployment mechanisms. Getting buyin from teams, ensuring _everyone_ learns the syntax.
And the unstated fear: The code it generates, is it actually good? Am I going to have silly issues down the road that are hard to debug and require diving into generated code to see some concurrency issue?
I guarantee you I won't be using your product. Just something to consider.
They failed. Full stop. There is no valid technical reason they couldn’t have had a smooth experience. There are numerous people with experience building these systems they could have hired and listened to. It isn’t a novel problem.
Here are the other companies that are peers that livestream just fine, ignoring traditional broadcasters:
- Google (YouTube live), millions of concurrent viewers
- Amazon (Thursday Night Football, Twitch), millions of concurrent viewers
- Apple (MLS)
NBC live streamed the Olympics in the US for tens of millions.
You know, like every other broadcaster, streaming platform, and company that does live content has been able to do.
Acting like this is a novel, hard problem that needs to be solved and we need to "upsell" it in tiers because Netflix is incompetent and live broadcasting hasn't been around for 80+ years is so fucking stupid.
I think you may mean easy. It may be _easy_, but it's not simple. There are so many more moving pieces, failure modes, operational issues now to consider. Websocket connections aren't free.
I choked reading this imagining people thinking they're doing something simple (as in not complex) by introducing websockets so they can keep state in their Go backend and sync it with their front-end, ya know, to keep track of the # of TODOs checked.