alcasa

That's not how this works. It really depends on how close you are to a position, where people might want to bribe you. Low provincial officials with direct ties to local land development might e.g. be able to take many more bribes than a highly ranked official in an office that is far removed from economic activity.

They should have used Figher Jet codenames instead. The MiG-15 one has a nice ring to it.

They forgot how to do pretraining.

I used it as a daily driver back in 2014/15 and it worked ok from what I remember.

Clinical research is in a weird spot, where you need both clinical experience and research experience. Getting the former already requires long hours and you are swamped with work. The latter is highly age gated and if you want to pursue research you often need to achieve specific milestones before a certain age, e.g. to be considered for tenure etc.

Isn't lisp getting advertisement every couple months here or so. After so many decades I don't believe the reason lisp doesn't have traction is just marketing.

You can still get your Radium Spa treatment in the Czech Republic: https://www.axxoshotels.com/radium-palace-spa

Employer cost is not 2x, more like 1.2x, employer overhead is mostly insurance related stuff. We had salaray to employer cost tables at my previous job.

What true though is that after taxes you might just receive 60% of your total salary once you deduct taxes and insurances.

I fear people will just get used to it. Nobody gets tailored clothing anyhmore and people don't question that we have standardized sizes that don't really fit anyone properly. People commonly buy standardized furniture and rarely get something to a specific for their room. If cheaper software (I mean thats mostly what it is) gets the job done, we will probably just keep doing that, even if that means we lose something in the process.

Obligatory Simpsons reference: https://www.youtube.com/watch?v=Pqb-VzkfRrY

Maybe controversial, but I believe a lot of OOP/Clean Code patterns are the software equivalent of corporate BS.

Frankly the most critical question is if they can really take shortcuts on DV etc, which are the main reasons nobody else tapes out new chips for every model. Note that their current architecture only allows some LORA-Adapter based fine-tuning, even a model with an updated cutoff date would require new masks etc. Which is kind of insane, but props to them if they can make it work.

From some announcements 2 years ago, it seems like they missed their initial schedule by a year, if that's indicative of anything.

For their hardware to make sense a couple of things would need to be true: 1. A model is good enough for a given usecase that there is no need to update/change it for 3-5 years. Note they need to redo their HW-Pipeline if even the weights change. 2. This application is also highly latency-sensitive and benefits from power efficiency. 3. That application is large enough in scale to warrant doing all this instead of running on last-gen hardware.

Maybe some edge-computing and non-civilian use-cases might fit that, but given the lifespan of models, I wonder if most companies wouldn't consider something like this too high-risk.

But maybe some non-text applications, like TTS, audio/video gen, might actually be a good fit.

Didn't know the L in Samuel L Jackson was for LeCun.

Fair enough. I feel like designing AI-proof take-homes is getting ever more futile. Given the questions need to be sufficiently low context to be human-doable in a short time and timespans for AI tasks increasing, I'm not sure take homes can actually serve any filtering function whatsoever, besides checking if applicants are willing to put in a minimal amount of effort.

No the 2 hours is their time limit for candidates. The thing is that you are allowed to use any non-human help for their take homes (open book), so if AI can solve it in below 2 hours, it's not very good at assessing the human.

Source for the 10B amount: https://www.cnbc.com/2026/01/14/cerebras-scores-openai-deal-...

Having worked in rare disease diagnostics in a non-US country with good public healthcare, most patients had to fight their way to the correct speciality to get their diagnosis. Without the persistence of family/specific doctors, its not possible.

AI might provide the most scalable way to give this level of access/quality to a much wider range of people. If we integrate it well and provide easy ways for doctors to interface with this type of systems, it should be much more scalable, as verification should be faster.

I think there might be a culture divide here. That person is very likely from Germany/Berlin based on their attitudes and descriptions and I feel like the hacker/tech scene is very different from bay area vibes.

FAANG is not really a thing here and people are much more tech-luddite, privacy paranoid.

I feel like the hacker response would be to roll your own models and move away from commercial offerings. Stuff like eleuther.ai is pretty inspirational, but that movements seems to have died down a bit. At least we still have a couple companies believing in doing open-weight stuff.

Tbf there are a lot of dishwashers, where I have had to essentially prewash all the dishes to make sure the dishes actually come out clean.

Why? Its 4.1m AUD. Given the salary of devs and the scope of such a website, the original budget seemed pretty optimistic.

There is a german idiom "Totgesagte leben länger" (the ones assumed dead, live longer).

Just because a business might be unsound, doesn't necessarily mean it will imminently collapse.

There is a lot of money going around right now and at its core Cerebras is a hardware company and their core innovation is in hardware.

At its core Cerebras is a hardware company. If they would have been able to just sell hardware that would be great, but for now I think the available speedups are not really worth it to most people to reengineer their software to fit the computation model outside of a few niche problems and TCO is pretty high.

But given the amount of resources being poured into AI right now, a lot of these tradeoffs don't seem so far out anymore. The main open question I guess is whether their architecture will be a good fit for whatever requirements come in the future (think continuous training and RL stuff). Their architecture is sufficiently different, so there might be some valid reasoning to hedge your bets.

A lot of people have fundamental beliefs different enough from yourself, that your torment nexus is just a normal good job for them.

You can take almost any tech job, and there is a probably an argument to be made, that it is mostly reasonable given some other moral priors.

Really cool, especially once 256k context size becomes available.

I think higher performance will be a key differentiator in AI tool quality from a user perspective, especially in use-cases where model quality is already sufficiently good for human-in-loop usage.

I honestly feel like there were some issues in their specific hardware selection. Linux shouldn't be that bad on a recent Thinkpad system

Or use it for an eGPU connector, I believe there were also options to use it as additional storage

If you mean Mac Pro and not Macbook Pro I wouldn't be that suprised. Single thread performance is mainly dominated by clock speed and possibly instruction set improvements. Having a much higher TDP there isn't much magic you can work to make a Laptop faster than a recentish desktop system.

Adguard, its also usable for free and doesn't have shady deals with ad companies.

> Great. I was waiting for React to stabilize before I learned it.

But once it's stable, it's outdated.