Hacker News new | past | comments | ask | show | jobs | submit | ptero's comments login

I may be wrong, but I think the alleged victim writes a financial column at one of the pretty big houses. And covered scams before. I would be really surprised if that person stuffs $50k in cash into a shoebox for an "undercover CIA agent".

Sorry if I confused this story with something else.

thanks, you got the story right! It does sound crazy, but that also seems to be the point: scam victims are shamed and it's worth level setting on how "it can happen to anyone" because there's very sophisticated psychological warfare going on here.

Why? An honest question, I thought undersea cable repair is a common, straightforward procedure.

In this case the repair ship might need some air defense, but US or UK could easily provide that.

Based on news coverage I've read:

- it's a specialized business requiring specialized ships which are often booked out a long time in advance

- the damage is in a war zone

Would you be willing to do your job in a war zone? Would your boss be okay with sending you there? With sending very specialized and expensive equipment there with you?

Also what do you think would be sufficient to provide peace time equivalent safety guarantees for the ship and crew? Military equipment and crew ain't cheap either.

They're actually not booked out a long time in advance. Usually in case of a fiber cut on a major subsea cable, the crews will be out to fix it within a couple days at max.

The real issue is the second fact you identified, that it's cut in middle of a warzone. I don't think they will risk sending a ship out unless closely escorted by military vessels.

It’s a high risk job even with naval escort. Risk drives up prices.

Welding undersea fiber under fire isn't in any of the technicians' job description, so while they say that everybody has a price, it's possible that the bid-ask is wide enough that you could park an Ever Given sideways there and there'd still be room. These folks probably make enough already.

So then it’s possible a US military option (involving the Army Corp of Engineers) will be required to fix it. Those soldiers have experience doing work in the line of fire and have no ability to refuse.

This. Good code is clean and has a well thought through internal architecture. LLM-ifying the code and treating it as a black box (if it passes the tests, it is acceptable) is tempting, but it works until it does not and the "does not" might come pretty quickly: once a human cannot easily untangle the logic the only fix is a rewrite.

I think there is a way to extend the useful life of such an approach by setting up a good architecture with lean, strict interfaces and thorough tests. Then one can treat any module that is compliant as a black box and give a computer the power to insert as much crap as it can generate. You then should be ready and willing to rewrite any box that has become so convoluted that LLM can no longer fix, likely by splitting it into smaller externally observable and testable elements.

I doubt that this is a long-term viable approach, but this is just a personal hunch. It would be interesting to see how such approaches develop. My 2c.

And he is saying AMD focusing on non-GPU markets and going from being a laughable underdog to a powerhouse hardly qualifies as "being asleep at the wheel".

Alright, fair point, but that is also why they missed out on the AI boom.

As if it was just there for the taking. AMD doesn't have the money to do what Nvidia does

Yes they do.

They were in the same ballpark until the past coupld of quarters. Nvidia laid the groundwork for this ages ago by developing drivers that play nice with NN.

AMD could easily have done the same at the time, and can do so now.

Even if it costs $1b-$10b to do it, AMD can afford it and it’s worth doing. They’ve got $55b in equity, and a huge $280b market cap to issue more stock into. [https:valustox.com/AMD]

I dunno, man.

They're releasing APUs with tensor cores right now.

They're finally getting their dogshit ROCm drivers fixed right now.

Yes, they're late. But they're definitely playing. This whole "asleep at the wheel" thing may have been true a year ago, but it's not true now.

[edit] Apologies, responded to the wrong comment, but I'll leave it here anyway.

Past couple of quarters? We're talking about the past. And their roadmaps go out five years. They have to spend now to get results in three to five years. What Nvidia was spending two years ago will largely determine how the next two years go.

Limiting the time of the protection would go a long way towards its acceptance. With the speed of technology today, 3-5 years seems reasonable. Anything beyond this is an overkill. This also discourages early filing (which now causes to mostly stifle related exploration) and encourages filing when the technology is ready for commercialization.

I also view the goal of the patent system as benefiting a society by encouraging the people to innovate. While somewhat similar to the current goals (rewarding the inventors) it is subtly different because its success and failure is determined not by the fairness to each inventor but by its statistical impact on innovation. If it tends to encourage innovation, even if in a few corner cases the time limits are too short for the inventor to reap the full reward, so be it. If it tends to stifle innovation by focusing on fully defending every inventor it needs to be repealed. My 2c.

A few of those thoughts I completely agree on: results of overhiring, and technology enabling bigger impact by fewer people seem spot on to me.

But a few I find strange. For example, a semi-implicit assumption that hiring, profits and stock price are tightly coupled. That is not always the case. They are correlated, but often have significant time lags, so it is possible for a company to hire rapidly on a great novel idea, realize a first mover advantage, have it lead to profits, have that lead to a stock rise by which time there is no longer a need for aggressive hiring. The company does not even need to do layoffs, just a stop to aggressive hiring is enough to feel a frozen job market.

Edited: apparently I totally misunderstood the meaning of the "non-regrettable attrition". Is the only difference with a layoff is firing employees one-at-a-time vs en masse?

> Is the only difference with a layoff is firing employees one-at-a-time vs en masse?

Theoretically, layoffs are not performance-related, although sometimes they are. Getting laid off doesn’t necessarily reflect poorly on the person who was laid off.

Firing is performance-related, and it’s usually a bad look.

The former may get severance and will probably be eligible for unemployment, while the latter will likely get neither.

Are you sure about the unemployment? Given that it is usually administered by the state (at least in the US) I am surprised that any non-criminal involuntarily termination would be rejected for the unemployment benefits. But IANAL and personally never needed one, so anything is possible.

> Are you sure about the unemployment?

Definitely not sure, and it probably varies based on location.

In CA, where I am, one criterion is “unemployed through no fault of [their] own”. Getting fired is widely considered to be the individual’s fault (e.g., non-performance). I’ve heard something about the standard being “egregious misconduct”, but that’s just beyond the details that I know about.

Getting laid off usually isn’t considered the employee’s fault (e.g., due to lack of work or change in direction).

Iirc, companies might have increased payments into the state unemployment coffers when they do layoffs versus firings. I have only heard about this second-hand, so I don’t know the details.

Anecdotally and perhaps only tangentially related, I have friends and family who have had to or chosen to fire people, and they have all said the same thing — layoffs are much easier than firings.

> Getting fired is widely considered to be the individual’s fault (e.g., non-performance). I’ve heard something about the standard being “egregious misconduct”, but that’s just beyond the details that I know about.

I have no real information, this is just my handwaving, but I do not think this is the case. Most employment in the US is at will for both sides. That is, an employer can terminate an employee whenever it is beneficial for the business (and an employee can leave with no warning even if he is needed by the business). No "fault" is required on the part of the employee -- needs and priorities of the business could have shifted, for example.

I think most employers would take this route of "no fault termination" if at all possible. Otherwise it would make sense for the employee to fight this, file complaints, etc. incurring potentially large costs for the company on time, effort and the PR fronts. So they tend to terminate employment with a bland corporate legalese instead of a red flag like "misconduct". Being terminated by a previous employer may cause some concern when job searching, but unless it actually comes with a paper trail of "gross misconduct" or some such it should not affect unemployment benefits. My 2c.

Yes, but hardware is hard. Reliably building up to enough volume to put a dent in Nvidia sales is extremely hard and independent of designing the successful product, tying it to current software APIs and convincing users. I am not surprised that a new company, however talented, cannot crack that nut. My 2c.

The velocity at which the field is moving dictates that you must not wait for AMD and Intel to catch up with a working and performant implementation, even if the TOPS per dollar are somewhat better on paper.

I would agree: while not impossible, it seems like some key information is missing.

On the one hand I would definitely expect zero empathy from a hedge fund. They should pay big bucks (pounds in this case), but if they consider an employee no longer useful it is an instant goodbye.

But firing a key engineer of a successful production system that they plan to operate -- no way!! Neither 48 hours after going live, nor 6 (or 12, 18 or 24) months later. Trading strategies always get refined, changed and tuned and an oops in that process can be very expensive. A hedge fund would fight to keep the key engineer who built the system.

One case where this is possible is if a bigger fund wants to snuff out a new entrant (for example because the smaller fund is messing up the larger fund's strategy). Then the owners get $$ and everyone else gets the boot. My 2c.

You generally cannot downvote old posts (24 hrs and above I think, but not sure), but can upvote for months (maybe forever). Was the post you checked with no downvote arrows older than 24 hrs?

Found another one to double check. It says "dead", so I assume that means it's at -4. It also doesn't say [flagged], so it's not flag killed. 19 hours old. No down arrow. No up arrow either, for that matter. No arrows at all. The only thing I could do was vouch for it.

But almost all of that author's recent comments are dead, so maybe what I'm seeing instead is someone who's shadowbanned.

Just a single data point and a personal opinion, take with a grain of salt.

The companies do hire principal and staff levels from such postings (as well as from other channels). But even more important, many companies are open to hire at a different level. They create a single posting (e.g., to avoid HR hassles) and if the person they like is at a different level they will adjust it in parallel with sending an offer. So unless it is clear that the company only needs principal level hires I would send a resume there as well.

For example, HR might give the hiring manager a single slot and insists on having a specific level in the posting. What this often leads to is the hiring manager aims as high as HR lets him (usually easier to hire at a lower level than approved than at the higher level). Those 2-3 levels might end up as a union of all requirements and skill lists can get ridiculous. Bottom line -- if the company is interesting and your skills are relevant I would apply.

The market today is not great. Many companies overhired during zero interest and covid stimulus times and are now reverting to the mean. IMO it will get worse before it gets better as unprofitable companies run out of money. So I would focus on areas where there should be less competition: secondary metros (not bay area); larger, profitable companies (expand beyond software-specific companies); hybrid or on-site jobs (fully remote are especially competitive now); etc.

If you apply, apply well: have a few polished resumes and, if needed a good cover letter (no need to be super company-specific; put company name, copy-paste 2-3 main relevant points from your database of snippets and let ChatGPT fix the style). Good luck!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact