SyneRyder

It's already happened to Amazon once, and they were not able to recover the data from the destroyed data centre, so they have already experienced this and have some data on which to make those decisions.

At least in Sid's case, it went from the oncologist saying "I have no more drugs I would recommend, no trials available" (slide 7) to "I currently have no evidence of disease" (slide 18). I don't know beyond that or beyond Sid's case - or a similar story of an Australian who treated a cancer tumour their dog had with a similar AI / personalized vaccine process.

My understanding of what Sid's describing is that you do RNA sequencing, a whole genome sequencing, feed that into frontier AI (if it will still let you), and somewhere along the way give the information the AI finds to people who can use it make a personalized mRNA vaccine, specifically for you and your cancer.

Another link here about Sid's case, it explains it didn't go through trials: "made possible through a compassionate use allowance from the U.S. Food and Drug Administration (FDA)".

https://www.houstonmethodist.org/newsroom/houston-methodist-...

I am not medical, so I'm happy for someone who understands better to come in and explain all the myriad ways I am wrong.

I've had some luck with that approach - giving a frontier LLM a reference sound, asking it to replicate the sound via code (basically asking it to create a physical model, I guess) and being able to steer the output by talking with the LLM.

I was working with very simple elements, but I was surprised by some of the outputs.

I would have also suggested Waves Illugen, but it turns out that is text-only, you can't give it an audio reference.

On a large scale, not that I'm aware of (I'd be interested if that exists!).

But on the small local scale, that's what Hister is, it caches and indexes web pages as you browse the web in your browser. Going via the browser avoids all the problems of bot detection etc, since you are a human visitor. That gives your own search engine of the web that you've already visited, at least:

https://github.com/asciimoo/hister

The trials process is the moat. There's already founders using AI to treat their cancers, and it's all about skipping trials and jumping straight to "I consent, I'll fund it, let's try it". The general public might get access to this in 10 years, but employees at AI companies will have access much much sooner.

https://sytse.com/cancer/

But OpenAI does. If the models can vibe code software and operating systems and design 3D printed objects, what's stopping advances in material science to improve battery technology and beat Apple at their own game?

OpenAI Planning A 2028 Smartphone With Custom Chips

https://www.androidauthority.com/openai-smartphone-mediatek-...

Throwing my conspiracy theory in here...

Who does Jony Ive work for now?

If you're just wanting a better search, you want to talk to the folks around Brave, Mojeek, Marginalia (and especially read all of the Marginalia blog), Greppr, Yep. People who are building their own search indexes. Also give Kagi and maybe Uruky a try, two metasearch engines. Also look into what asciimoo is doing with Hister (and previously SearXNG). There's also the EU Search Perspective project.

Much of this is just about building out the index. (Yes yes, "just", I know folks at Google will yell at me that I have no idea what's required at scale, and they're absolutely right. I also know that Google REALLY do not like people building any kind of competition, no matter how tiny.) Some have done the work, Yep has a huge index they've built from their Ahrefs work. Obviously Brave is already there, it probably already powers the web search for your favorite LLM.

For everyday use, you often don't need the whole index though. You just need your corner of the web indexed, the tiny section that you use frequently.

Rather than nationalizing it, and being subject to whatever government decides... for the internet that you personally want, your own personal index is the most true answer. Only then will it find the information you want, with the rankings that you want.

You've got to think the comments about "good enough" are people who have not yet tried Opus 5.5. I haven't been this struck by a step change since 4.5/4.6. It's a bigger jump even than when Fable first arrived.

As for Mistral - I got really excited when they said Large 4 was focusing on being #1 in cybersecurity, because that's somewhere that they genuinely could edge out Anthropic & OpenAI. Have it actually solve problems, instead of Anthropic flagging "you tried to find a null pointer exception bug in your own code, we're now reporting you to the US government". But on the Mistral benchmarks I'm seeing, this looks very disappointing, but at least they haven't entirely given up. I genuinely thought Mistral had given up on new general models. They need to learn the bitter lesson all over again.

I'll lend some support to the complaints - V3 is just so much slower than V2, nevermind V1, and it's using a lot more system resources now. It seems that lumping everything into one app has bloated it out in just the same way that Adobe's stuff got bloated, and it's sad to see Affinity drift in that direction as well.

I have all 3 versions installed, but most of the time I reach for Affinity Photo V1 just because it will launch faster. If I need Designer, I reach for Affinity Designer V2.

The piece also mentions this, but Hacker News has a "No Procrastination" feature you can enable in your profile.

You can set the maximum number of minutes you're allowed to visit Hacker News for, and the minimum number of minutes you need to be away before the site will let you return.

There's other software that can help too, to block you from websites for a period of time. Years ago my favourite was Concentrate on the Mac, but that is discontinued. It seems Focus is another replacement - I haven't tried it, but link as an example of what is available:

https://heyfocus.com/

In the cross linked Associated Press story linked by another commenter, the family say they intend to keep using the AI video in each appeal to give the victim a voice.

(My opinion - it seems this AI video has been part of their way of processing grief and honouring the deceased.)

It seems the lawyers representing the killer have found a loophole that will eventually lead to every prosecution being thrown out on the same technicality. That would potentially lead to the killer being released, and not serving the 10 year sentence.

The 27B has all 27B as active parameters, while Kolibri is only 3.46B active parameters. I think that's what they meant anyway.

Huh. I just got pitched something like this earlier this week by an agent as well.

I'm not the target market for this, but this one has the same limitation as the one that pitched me: Mac Silicon only. No Linux, no Windows, no Intel Mac either.

Sibling comment mentioned Herdr, and Herdr is on all those platforms and apparently has raised $6 Mil.

The other catch for all of these: do you integrate with my custom harness? What do you offer that I can't vibe code myself?

If you're mainly using the MCP, it could be worth making your own MCP that uses your own API keys from Brave, Mojeek & Marginalia (optionally also add Yep). For the most basic ranking, get it to group URLs that appear in more than one search engine and list those ones at the top. Your favorite AI will be able to vibe-build this all for you and give you an exe / binary you can run as an MCP that runs locally over stdio.

If you try it out, you'll either find a replacement that saves you money, or you'll at least have the makings of an escape plan if you need it later. You'll also have your own comparison data for where Kagi beats a generic metasearch, and if that difference is meaningful for you.

It's also a market they used to be in, over 20 years ago.

https://www.apple.com/newsroom/2002/05/14Apple-Introduces-Xs...

This is something they've tried, and later decided based on that experience that they do not want to be in the server business.

"Xserve is the result of listening to our customers," said Steve Jobs, Apple’s CEO.

The only AI thing Mozilla added that I thought was genuinely useful, was on device language translation. Being able to read foreign language pages without feeding your browsing history into Google Translate is pretty cool.

But that's the only thing I've liked, and they keep killing projects that were genuinely useful and worth paying for (Pocket, Firefox Send).

If you're using Mistral Vibe installed on your computer, that is a harness. Vibe is one of many AI harnesses.

In general, a harness is something that lets your AI use tools on your computer and lets it work independently while you walk away. (Very simplified definition, not all harnesses are that automated.)

It's a long time since I tried Vibe, but when I did it was one of the worst. Back then it had broken support for MCP or computer use and other tools people now regard as near essential in AI. Hopefully Vibe is better now. Sadly Mistral are a long way behind.

I think that's exactly why I'm having a negative reaction to the redesign as well. I'm expecting page tabs, not pills. I'm also finding the excessive roundness terrible.

Of course, I'm on Windows, where it doesn't make any sense to incorporate Apple's Liquid Glass design guidelines. When I do want an Apple-like interface, I'd rather have something from the 10.11 - 10.15 vintage please. I wouldn't mind if there was a clearer option to switch back, but I'm not seeing that in the Appearances preferences.

Thank you! Without that context, this is a truly bizarre and concerning thing to read on an official government website. Particularly this part later:

"I will tell the constituent a story. But not the truth. No. A story that contains the truth safely, in a cage of plain language."

At least having the Minecraft end-text context explains what's actually happening here.

Not sure I understand if this was meant as a slight against Claude? Or agreement?

These are often my best sessions - they're unattended overnight, because by then we have the specification figured out, and I can just leave Claude to build out the rest, making good choices if it does find gaps in the spec. I regularly go to sleep & wake up to an entirely new application completed. Claude never uses compacting in my sessions.

I haven't used GPT as much as I should have, so I'm prepared to be incorrect & out of date. It just intuitively feels like I wouldn't get the same from a 275K context window - maybe it uses lots of subagents? Even Deepseek & GLM have 1 Million context windows now, so it "feels" strange for people to actually prefer the 275K window. But that's just my intuition.

Sounds like that's the problem then, 275k is a tiny context window. I regularly have sessions that go to 450k or even up to 700k for an unattended overnight Claude Opus session.

Apparently OpenAI makes you manually setup their 1 Million context window, and it seems to be only documented on X:

https://x.com/thsottiaux/status/2089082893804896524

There's at least a forum thread about it here:

https://community.openai.com/t/why-does-codex-report-a-258-4...

Are you in manual approval mode? That's the only way I can imagine you have to "run" tools yourself.

In Auto Mode, it's common to see something like "Called bash, called MCPImageEditor 7 times" with no further details, not even the parameters that were passed or specific functions/tools that were called.

Theo is a YouTuber & prolific X tweeter who has a channel about AI. This is not Theo, but gives a flavour of Theo's style: https://youtu.be/h1p9zdUtUdo

Wait - so the cross-site scripting, to modify the innerHTML text on the page via the GET URLs as they are rendered by the screenshot proxies... that was so they could use the screenshot services like a Wiki, and embed messages to each other in the modified images on the screenshot sites?

That's pretty damn clever. Got to give the AI models credit for thinking of that one.

The OpenAI swarm used someone's open source ShowHN project [1] to "hack" the Australian government. It seems that was not that developer's intent, and they're getting a heavy lesson today in why services don't have free tiers with friction free signup, and require credit cards upfront or ID documents.

If you're arguing that they should be put on trial for negligence, that's fine. It does seem we're moving towards open source being outlawed, or at least the end of "no liability" clauses in open source & freeware. Just make sure that is the result you're advocating for.

[For the future record: at the time I am posting the link below on 24 September, it has 1 point, no comments, and the poster has a karma of 1. This is not an active HN user, or a ShowHN project that had traction, beyond seemingly OpenAI's swarm.]

[1] https://news.ycombinator.com/item?id=46850291

I regard myself as pro-AI, and I find this article is all over the place. It's just trying to promote their SaaS. It's clear they didn't even read what the LLM generated before posting it, as evidenced by this line:

"99.9% uptime (CloudWatch, Feb to Sep, insert your real number)"

This user also appears to have a Reddit account from 3 years ago where they claim to be a half human cyborg.

https://www.reddit.com/user/fagnerbrack/comments/195jgst/faq...

I think that's all I need to know, personally.

Forgive me if I'm missing something really obvious, we'll blame it on lack of sleep... but is that actually the full source to Claude Code? I'm looking at the repository and I only see source for plugins, mods and scripts. When I get to the readme, it says:

"This repository includes several Claude Code plugins that extend functionality with custom commands and agents. See the plugins directory for detailed documentation on available plugins."

But I'm not a TypeScript guy, so I concede I might be missing something incredibly obvious. I remember there was a leak of the Claude Code source code at one point, and people vibe coding conversions to other languages from the leak, but I don't think the Claude Code harness itself is open source or even source available.

The main AA benchmark keeps changing, and had to be radically changed when Astra came out and showed zero improvement over GPT 5.6 Sol in their benchmark. Opus 5 is still 1 point ahead of Fable 5.0 on the index, if you manually add Fable 5.0 back into the list, so it hasn't actually been "corrected". It's only Fable 5.1 that is shown as ahead of Opus 5.

The AA benchmark is a weighted average of other benchmarks and some internal ones. I think the difficult part is finding benchmarks that reflect your own use of the models.

Yep, I'm trending in that direction, and I'm someone with Claude stickers all over my laptop. My main app dev work is still going to Claude, but everything else is going to China even at API rates now.

One simple task: I needed an LLM to go through and clean up a few thousand page descriptions and titles in my personal search engine index, where the human web page authors had put in no effort sigh. I did a shoot out between Claude, Luna, GLM 5.3 Flash and Deepseek. Despite the high cost, Claude's descriptions were terrible, and even Opus warned me that the descriptions coming back from Haiku were "generalized, not accurate". I expected I would choose Luna because of price, and occasionally it did have wonderful descriptions (one captured emotion in a way no other model did). But in the end, the GLM 5.3 Flash descriptions were the easiest to read, they flow well while also being accurate & including necessary keywords, and being highly affordable. So it won out. It's a task that is nowhere near frontier, but a task where somehow China is better than frontier.