News for hackers. Stuff that matters.

londons_explore

Re:Astra can read and write weights to improve Qwen on Tuesday November 03, 2026 @04:46AM (#50349952)
I think it would be worth a human attempting this task....

On the face of it, running a model to a specific state, looking where is makes a mistake, and then figuring out which weights need to be jacked up to get the desired outcome doesn't seem like an impossible task...

I forsee this getting fixed by inflation, capital gains tax to stop people avoiding an inflation haircut, and an unrealised gains tax to stop people deferring paying.
Re:In LLM era, can software startup survive? on Monday November 02, 2026 @02:29AM (#50334047)
I find myself building far more simpler single use tools that do exactly what I need than using commercial swiss army knives which usually overcomplicate what I need to do.
The closest thing in my life to these zinc cable sockets are the zinc nubbins on the end of steel cables for bicycle gear and brake cables.

And I've seen a worrying number of those fail after decades too. In cases I've seen, the fact that load is uneven between the strands usually seems to play a part - some always start to snap or pull out before others.

Re:Bidirectional Charging During Grid Outages on Sunday November 01, 2026 @01:41PM (#50328382)
Likewise, there are plenty of plug-in hybrid cars which ought to be great home emergency backup generators (ie. they have a 40+ kilowatt generator and 7+ kilowatt inverter), but very few have this functionality even if the hardware in most cases supports it and would simply require a firmware update.

I had good luck with telling claude to rewrite my laptop bios to enable various power modes which were theoretically possible but disabled by firmware, like powering a fullsize screen from the laptop battery over usbc.

Maybe I'll tell claude to rewrite the car firmware too...

Re:Serbia just elected the world's brainiest parliament on Saturday October 31, 2026 @06:42AM (#50314767)
Whilst it seems like a good idea to have academics run government, history has shown that countries run by engineers and scientists tend to end up with some great roads and dams, but failing economically and socially in almost every other way.
Kinda sad that LLM's let me build little projects that would take a week in an hour, yet Google still isn't releasing much new or different, or even fixing bugs faster.

Mr boss man needs to uncork Google's 'labs' culture.

Re:Binance-hacker-1029 – $5.6B stolen from Binance on Friday October 30, 2026 @04:44AM (#50301277)
> access "in case they need it"

A good solution to this is to have a "break glass" system. Ie. access permissions get granted to the developer directly in an emergency by going through a set of steps available beforehand. All requests by that developer are then allowed, limited to 10 kilobytes of results per minute and logged.

For example, imagine you run a bank, and a rare code bug means that at exactly midnight on new years eve all bank balances have $1B added to them.

Your oncall takes down the production servers to to reduce losses and uses the break glass system, by getting approval from n other developers, allowing them direct database access. They use that access to count(accounts) with a balance around $1B or more, see it is everyone, and subtract $1B from them all, before bringing production back up with an older software release. Downtime = 2 hours.

In the morning, all the developers actions are reviewed.

If the developer turned evil and trying to dump the database, the n approvers and 10 kilobytes per minute would reduce data loss. But the break glass system still lets the developer make fairly major emergency fixes to reduce losses or downtime.

Re:Binance-hacker-1029 – $5.6B stolen from Binance on Friday October 30, 2026 @03:31AM (#50300799)
So when tether is blacklisted, is it reissued?
Re:Binance-hacker-1029 – $5.6B stolen from Binance on Thursday October 29, 2026 @06:34PM (#50297183)
Only down 3% today, which is pretty much nothing for something that regularly sees 20% swings
Re:Gemini 4 Carbon on Thursday October 29, 2026 @03:27AM (#50285546)
I dunno... DeepSeek has much lower prices, yet still demonstrated substantial profitable inference compared to hardware rental costs.

It all depends how many billions are going to paying salaries...

Re:Gemini 4 Carbon on Wednesday October 28, 2026 @02:44PM (#50278899)
When every day matters to profitability, I'm really surprised providers do voluntary cyber security testing or government approval in the critical path.

To be honest, if I were an LLM provider I would be wiring up my release process to be fully automated so within 10 mins of the last bit of fine-tuning it got released and a blog post posted.

Re:Early progress on accelerating drug development with Claude on Wednesday October 28, 2026 @06:06AM (#50271239)
I think the next big step is automated fast labs.

The equivalent of a GPU but for bio lab experiments. It should be able to do all the basic lab operations of mixing chemicals, heating, cooling, various tests etc tens of thousands of times per second.

Then you hook many AI's up to the automated lab and watch science happen far faster.

I don't think you can look at any lease system and claim success until leases have expired and been reissued many times and the system achieved steady state.

That isn't the case in Shenzhen.

Re:Anthropic S-1 on Tuesday October 27, 2026 @10:02AM (#50259537)
Are they really commitments if either side can cancel on the other after 4 months?
Doubt you'd be storing important ssh keys on the same VM
I really don't understand why Microsoft didn't build this when they released 64 bit windows.

16 bit windows wasn't complicated, and a simple VM and API translation layer could have kept compatibility going back 30+ years for not many engineering hours.

Yet instead they dropped 16 bit windows support and probably spent more engineering hours to support all the ancient software trying to update, and users who held back upgrading windows because they needed said ancient software, etc

Re:Diamond Transistor on Tuesday October 27, 2026 @03:20AM (#50255631)
Sure, but big ones still are expensive compared to say sand or other hard materials.
Re:Diamond Transistor on Monday October 26, 2026 @01:18PM (#50247386)
Lab diamonds are used in huge quantities (ie. By the bucketful ) for making cutting tools like saws and drill bits.

I believe the main difficulty is making large and pure diamonds. Making small low purity diamonds is cheap and easy to scale.

> but held off reporting details until the data could be presented at an upcoming scientific conference

If you hold the key to saving a million people's lives, yet you decide to let those million people die so you can take the stage at a prestigious conference, are you a good person?

Curing disease is literally a race, with people dying every minute of every day. In any other situation, labs would be running night and weekend shifts and not a second would be lost.

Re:Show HN: Steal My Bitcoin on Sunday October 25, 2026 @11:15AM (#50232451)
Yes, just the auditor looked like she was ready to call in the police for fraud when I said it.

Also, I live in a place where that was probably many months pay to the auditor.

Re:Show HN: Steal My Bitcoin on Sunday October 25, 2026 @04:53AM (#50229548)
I sometimes posted a donation link or joke site and got surprisingly large donations.

Made my tax audit rather painful.... "I posted a joke on the internet about binary numbers not being worth anything and someone gave me $10k of bitcoin that I'm now declaring - is it income for the work of writing the joke, or a gift? The sender likely didn't know who I was, and was likely overseas but I cannot prove who they were or which country they were in".

> Take the unexpected resurgence in book sales

At least some of this will be AI companies buying and scanning every book available...

Presumably that's why there are also lots of part time contractors.
If they asked harder for donations, they would surely raise enough for maintenance.

Plenty of companies can donate 0.1 engineers to work on php, if for no other reason than to make php better for their own use cases.

So these messages are presumably encrypted with a one time pad, and some lookup book of what each message of just a few characters means?

And I assume there are also a high bandwidth comms channels, and this is just the backup which is presumably tested each day.

The next version is sure to have signed firmware which sets these configurations to prevent this sort of thing.
Re:Kimi K3.1 Pro and K3.1 Pro Highspeed on Saturday October 24, 2026 @06:13AM (#50219787)
Worth noting the Claude code client it seems has been designed to make it rather difficult to start a conversation with one provider and then continue it with another.

Presumably they don't want you leaking your conversion transcript to a competitor.

Re:Kimi K3.1 Pro and K3.1 Pro Highspeed on Saturday October 24, 2026 @06:11AM (#50219774)
If you use the API, there is nothing stopping you writing both sides of the conversation...
CT logs stapling proof, but then never actually including that cert in the log.

Older »