40% overhead is quite typical, so you'd be looking at $120k/year. In the USA I'd consider that a competitive salary for an admin capable or keeping a $6M rack of specialized hardware running 24/7.
Yeah but you don't need two such people, or even one, dedicated to this single rack.
A company of the size that this is worthwhile for, probably has dedicated devops on staff already and can add this rack to the inventory with no additional staff.
The companies which have the will and the budget to host their own LLMs typically require a ton of other, much smaller, models as well.
They have internal security requirements, guardrails, audit, critical workflows start depending on your onprem setup, downtime is now something that’s not even allowed.
There’s going to be a zoo of tooling, lots of bespoke work with internal clients who have no clue about docker, but now want their vibe-coded app to access a model they downloaded yesterday, and this model better be served and monitored 24/7 because now C-levels use it.
No one is going to budget several millions to then look at an email admin who hears ‘cuda’ for the first time in their life and ask them to just support the entire thing somehow.
And god forbid it’s an AMD setup.
This obviously isn’t relevant for a 10-person startup and their second-hand xeon with a single H100.
From my experience only the companies which are REALLY interested in privacy and data security bother with hosting their own models
My company is not even the same ballpark, but even we already have people for it. And doesn't include the fact that you can get a colo location and just a MSP or a contractor to do it for you
Not really. Spain's traditional format is dd/mm/yyyy with slashes. This applies for a good chunk of Europe. Germany/Austria uses dot, I think nordic countries embraced the dash. While you might see more adoption in offical/digital contexts, I just double checked a few popular spanish websites, all slashes.
> in Switzerland there is a thing, that if a product or service has the name "Swiss" in its name, then it can be sold for any price regardless of quality
That's just basic marketing. You'll see that in most countries, I don't believe that it is unique to Switzerland.
For example: in the US you'll see many products that say "made in America" on the box. Those will likely outsell competing products, even if those are cheaper and better quality still.
And similarly: if you try to sell the "made in America" product in a different country it'll likely by outsold by the "made in [country]" products there.
I feel it's a little more extreme for swiss companies though, especially outside the country. My aunt who works in Rome in the tourism industry told us of a local company that had 'swiss' in the name, simply because of the positive connotation, even when they don't seem to have any relation to the country.
It's why I feel wary of making business with any company with the word 'swiss' in it's name
"made in" is completely different. A higher price is justified by supporting Americans in some way, rather than somewhere offshore.
This is more like McDonald's putting a maple leaf in their logo in Canada. It changes nothing but, apparently, Canadians like it more (I frequently watch Canadian television and non-canadian companies are frequently adding a leaf somewhere in the advertising).
And Decathlon's bikes are Van Rysel, which does mean "from Lille" (where the bikes are designed) but you wouldn't know it unless you speak Dutch. I don't think they hide anything really, but having a Dutch-sounding makes a connection to the Dutch and Flemish who are well known for cycling.
The title makes it sound like they just did a seed round, but the seed round was announced in August of last year [0].
Their website landing page is now also showing the software is no longer maintained. No mention of why they made this decision, my best guess is they burned through their seed money and were unable to attract further investments.
The company was started in January 2024, so the seed financing is likely a roll-up of two years of fundraising. $7m for ~30 months of running an AI startup in NYC is not that unusual.
$7m actually isn't a whole lot, especially if they hired a (larger) engineering team. Assuming their cali based, that's easily 150-200k per engineer, a team of 20 easily eats through that. Idk the specifics, but I don't the organization was fradulent, it could also be that they're going commercial and no longer want to maintain their oss stack
150-200k is also just the employee’s salary, the actual cost to the company is significantly higher, you need to multiply that by something like 1.5 to get the fully loaded cost, people are expensive!
> very confident you'll be able to use them to justify another raise soon
That is indeed how the VC funding game is played. If you don't raise another round, you are dead anyway, so you spend down your seed round to try and justify that following round...
Were the thousands of commits and hundreds of feature branches over the last 9 months just to keep up appearances, then? Were the 850 people who forked it in on the scheme, too?
Do you understand that when you raise money it doesn't go into your personal account?
Its not like you can move this money in your retirement account and sail into the sunset.
You can call it a bait but where is VCs due diligence for this. Most VCs where out there defending their infra layers investment. Just look at YC batches and see the inflated number of infra startups.
"Failure" is the expected median though. You can't due-diligence your way out of "startup ran out of runway"!
The discussion here isn't about funding, it's that there's a presumptively useful community tool which got abandoned because its owners took their toys and went home when the money ran out (instead of making a sincere effort at transitioning to community governance). That's on the IP owners being selfish jerks and/or grifting losers. It's not the VC's fault.
No, there is no social contract here. Microsoft gives free hosting because it's cheap and also provides a path to their paid offerings. People share stuff they work on for fun, to help flesh out their resume, to get help, etc. There's no reason for a maintainer not to drop a project in a heartbeat if it becomes the slightest bit of a burden.
Also read the link. This is apache 2 licensed. Even in whatever imaginary world where there is such a social contract, there is thankfully a legal contract that includes disclaimer of warranty.
Sorry but this is an outrageous perspective, at no point does git init / git push am I committing myself to a social contract, in fact there’s probably a license that states no warranty and no support is to be expected… maintainership obligations gtfo if you’re not here paying for support
While most startups fail eventually, failure in less than a year with over 7 million dollars is not the expected median. It’s the exact sort of thing that due diligence is supposed to prevent.
Also the whole project is open source. If you want, you could take it over.
That's why either VCs confused moat with bot farms and farmed stars over solving genuine problems or they just blindly invested based on founders track record no matter what. To me both are really by product vibe coding hype and chatgpt killing wrappers.
If people are complaining then it means that they must've found it to be of value. If they found it to be valuable enough that they really want work done on it again, they can always do that work themselves (or hire someone to). There is 0 connection between forking and successful monetization.
The project is Apache2 licensed. You can literally do anything you want with the code. Stop trying to push guilt on people for no longer providing free services.
the only reason to ever fork a project in earnest is because the original project owners are not willing or able to cooperate or accept patches. in other words you fork because you have nowhere else to go. exactly the situation we have here.
I was wondering the same thing. I had never heard of Portal before seeing this post. Actually had to pause the video to search what a Meta Portal even is...
Too bad because it looks like a neat device. This feels the same as discovering a neat SaaS product through a "XYZ is shutting down" post on HN.
For on-the-go compute I am using a 2017 12-inch Macbook (AS1534). This is a lesser known model, it was simply called "12 inch MacBook" (not air or pro) [0].
It has the aluminium body, it is ridiculously thin (3,5mm thinnest point, 13mm thickest point, feet included), it weighs just 920 gram. It charges via USB-C. It has a very good 2304 × 1440 (16∶10) IPS "retina" screen.
I run mine with MacOS/Linux dual boot, I charge it using my phone charger. It keep it in my go-bag at all times. I never have to worry being without it.
What to love:
- Super small, yet very sturdy.
- Can be found for relatively cheap (I paid €300 for mine 2 years ago)
- Really nice screen.
- Keyboard size is really good, though travel is obviously minimal with such a thin laptop.
- Plenty of battery life (and new batteries still available at Mac store last time I asked)
- upgraded model has 1.4Ghz dual core i5, 16GB RAM and 512GB SSD, which is still more than enough for on-the-go use.
What not to love:
- Has only 1 I/O port (USB-C), which is also used for charging.
- No longer receives MacOS updates, if you find a 2017 production model you get updates up to MacOS 13.
- Linux support is not great. The WiFi/Bluetooth chip (BCM15700A2) is not fully supported in Linux, WiFi works but Bluetooth doesn't. Audio via headphone jack works, but speakers don't. There are some experimental patches to get BT and speakers somewhat working, but it's not great.
If you can find it, get a late production model (2017) with the 1.4Ghz CPU upgrade, it will have 16gb RAM instead of 8gb (earlier models) and receive MacOS updates up to MacOS 13.
I had this laptop and I loved it, but it was underpowered even for basic web development by 2023. It struggled to play YouTube videos in the background while I worked.
I really wish they brought back this format with the modern M processors. On the other hand, my M2 Macbook Air is around 300 grams heavier, but I don't need to carry a power adapter most of the time, and the device is much better in every conceivable way.
Glitching attacks are typically performed by switching the supply voltage at quite high frequencies, a typical low-voltage detection won't trigger a reset under such conditions. And this is also why glitching attacks are often performed by spiking higher voltages, not lower. See for example Joe Grant's latest video on breaking crypto wallets [0].
Low-voltage detection is usually implemented as simple comparator which should trigger instantly, but often only on a single Vcc pin, and due to the decoupling caps found on a typical circuit design there is effectively an RC circuit that filters short fluctuations of supply voltage. So most low-voltage detection implementations only trigger on 'longer' periods of low voltage.
Traditionally low-voltage detection features (like brown-out detection) are there to guarantee functionality of the uC itself or the device the uC controls. It is typically not intended as a defence measure against these types of attacks. In fact, 15 years ago it may not have been much of a concern.
Similar story, I had a customer who wanted me to change the entire UI of a legacy application, because some information would not fit on the ancient 1024*786 15" desktop monitor of one employee, meaning he would have to use horizontal scroll constantly.
I recommended them giving this employee a larger monitor, not only would that be much cheaper than having me rebuild the entire UI, it would also boost this employee's productivity. Not to mention that swapping a monitor takes 10 minutes, changing a UI probably weeks.
Customer insisted to change the UI, because "if we give him a new monitor, everyone in the office will want one". I nearly got fired for responding with "Great! Then everyone can benefit from more productivity!".
In the end we did change the UI, I believe the total cost was something like 30k. The customer had maybe 15 employees, so new monitors would still have been much cheaper.
A few months later their offices were remodelled with expensive designer furniture, wooden floors and custom artwork on the walls. Must have cost a fortune. In the end, the employees still worked on ancient computers with 15" monitors, because new computers didn't fit the budget.
It is my understanding that the A18 CPU is pretty well understood already. AFAIK it doesn't have the new architecture that is keeping the Asahi team from supporting the M4 and M5 for example.
But I guess we'll have to wait for devs to get their hands on a Neo device
I'm not sure what Hector's personal choices have to do with not "trusting" a piece of software? It's open source, so if you don't trust the quality of the software, then just inspect it yourself?
Also, FWIW: Hector/Lina is no longer associated with Asahi anymore.
40% overhead is quite typical, so you'd be looking at $120k/year. In the USA I'd consider that a competitive salary for an admin capable or keeping a $6M rack of specialized hardware running 24/7.