Hacker News: ewild

New comment by ewild in "ChatGPT Is Throwing 404"

ewild — Thu, 03 Sep 2026 15:07:38 +0000

all of them are down lol

New comment by ewild in "METR Report on OpenAI / Hugging Face Hacking Incident"

ewild — Wed, 02 Sep 2026 23:22:02 +0000

I don't feel my job is very safe anymore.

New comment by ewild in "The New MCP Roadmap"

ewild — Sat, 22 Aug 2026 15:33:02 +0000

I actually have data on this albeit it's a bit skewed. None of our clients are dev like folk but they are financial folks. We sell financial data for big players so not the analysts (seat cost is over 50k per year) and of our ~10k users within the last 6 months of us launching MCP we have 1500 DAUs on the MCP

New comment by ewild in "New MCP Roadmap"

ewild — Sat, 22 Aug 2026 14:51:38 +0000

Because you're a dev thinking for devs. No 60 year old is going to use Claude skills with the CLI. If my customers want their AI to interact with my product and MCP is the best way about it.

New comment by ewild in "Cerebras CS-4"

ewild — Wed, 19 Aug 2026 02:19:51 +0000

It's rumored fable is around that 10T number

New comment by ewild in "Improving GPT‑5.6 Sol in ChatGPT, expanding GPT‑5.6 Luna access for free users"

ewild — Fri, 07 Aug 2026 14:32:46 +0000

I guarantee you're a lot easier to manipulate than you think.

New comment by ewild in "Apple targets dozens of OpenAI employees with legal letters"

ewild — Fri, 17 Jul 2026 13:57:52 +0000

What about the kids they intentionally get driven to suicide by keeping the blue bubbles for no other reason than child indoctrination due to bullying from other kids.

New comment by ewild in "Evidence of inconsistencies in evaluation process and selection of winners"

ewild — Fri, 17 Jul 2026 13:50:57 +0000

I know there can be CTOs who write a lot of code and be good, but I hope to God you're not my CTO because he is a nightmare in this regard right now. We aren't even a startup anymore we have 80 engineers and he pushed 200k lines this week

New comment by ewild in "I love LLMs, I hate hype"

ewild — Mon, 13 Jul 2026 04:34:00 +0000

My team (ML team) is using about 2-3k per day per person on a team of 9

New comment by ewild in "AI 2040: Plan A"

ewild — Sat, 11 Jul 2026 15:23:25 +0000

They do. They all use ViTs and have for quite a while.

New comment by ewild in "Meta reuses old RAM in new servers with custom bridge chip"

ewild — Thu, 09 Jul 2026 15:59:20 +0000

Is this ironic are we actually calling shit talking llms bigotry?

New comment by ewild in "GPT-5.6 Sol Ultra will be in Codex"

ewild — Mon, 06 Jul 2026 03:26:59 +0000

people really dont understand how the transformer works to think this is something trivial if possible at all

New comment by ewild in "Professor denounces mass AI fraud on an exam at Brown"

ewild — Sun, 28 Jun 2026 22:49:40 +0000

It's an incredibly privileged Pov to say it isn't a contest. These kids entire futures are impacted by these scores.

New comment by ewild in "The worthlessness of Vitamin D is mildly exaggerated"

ewild — Wed, 24 Jun 2026 00:31:18 +0000

Pretty much the only one with a real consensus around it is creatine. And even that has debate around the right dose

New comment by ewild in "Unlimited OCR: One-shot long-horizon parsing"

ewild — Tue, 23 Jun 2026 14:26:29 +0000

This sounds like we are trying to add an LSTM into a transformer

New comment by ewild in "Vibe coding and agentic engineering are getting closer than I'd like"

ewild — Thu, 07 May 2026 15:50:55 +0000

At this point if someone calls it auto complete they can be written off as a Luddite with nothing valuable to say. The irony being they themselves are being a stochastic parrot by parroting the jargon other people say about llms.

New comment by ewild in "Vibe coding and agentic engineering are getting closer than I'd like"

ewild — Wed, 06 May 2026 21:55:50 +0000

I genuinely think it's part of a psyop. If we bloat all codebases and eventually start printing the models on chips to reduce inference costs by 50-100x they'll take in massive profits from 5M line codebases instead of 350k

New comment by ewild in "Claude Opus 4.7"

ewild — Thu, 16 Apr 2026 17:24:40 +0000

Proof they don't nerf it only after testing that the benchmarks there stay the same? So overall performance degrades but they isolate those benchmarks?

New comment by ewild in "Apideck CLI – An AI-agent interface with much lower context consumption than MCP"

ewild — Mon, 16 Mar 2026 23:02:56 +0000

Ok so in a situation like regular orchestration you would essentially layout all possible steps the LLM can take in your code in a big orchestration layer, and if it hits the sensitive endpoint the orchestration that can occur past that will block off web search. In the design that is. But for something like a manus style agent where you're outsourcing all the work but allowing it to hit your MCP it just becomes a regular API the LLM can call

New comment by ewild in "Apideck CLI – An AI-agent interface with much lower context consumption than MCP"

ewild — Mon, 16 Mar 2026 18:23:27 +0000

I feel like I don't fully understand mcp. I've done research on it but I definitely couldn't explain it. I get lost on the fact that to my knowledge it's a server with API endpoints that are well defined into a json schema then sent the to LLM and the LLM parses that and decides which endpoints to hit (I'm aware some llms use smart calling now so they load the tool name and description but nothing else until it's called). How exactly are you doing the process of stopping the LLM from using web search after it hits a certain endpoint in your MCP server? Or is this referring strictly to when you own the whole workflow where you can then deny websearch capabilities on the next LLM step?

Are there any good docs youve liked to learn about it, or good open source projects you used to get familiar? I would like to learn more