HNHacker News
TopNewBestAskShowJobs

brutus1213

1,242 karma · joined November 8, 2016

submissionscomments
brutus1213··on Canada’s new tech talent strategy
I already bought a home (it is a starter home) and it is nice seeing my home price increase but I can never upgrade (well .. I can but it would be financial suicide to). I think some Canadian home owners get that but your point is actually totally valid. The Home Affordability issue just targets renters.

I think the right billboard should cover our shit roads and transit, and our healthcare/ER. That affects everyone (and seniors the most!)

brutus1213··on Canada’s new tech talent strategy
Issue is not just housing. Our health services and infra is crumbling. 401 near Yonge st is jammed nearly 24 hours (yes!!!). We literally have no more space for cars. We have switch to the 407 for as much travel as we can (and it is crazy expensive and silly given how much tax I pay as a techie).

The ERs in GTA are just always clogged. I've had to deal with multiple instances of 10+ hour waits in the ER. When they were not addressed in time, one such incident got worse, and ended up in a 10 day serious hospital stay (love the logic on that one).

Getting an appointment with any sort of specialist is also a gauntlet.

I'm an immigrant and it was surreal hearing the new Toronto mayor talk about how much she accomplished as an immigrant to Canada. It was indeed possible when she and I came to this country. It is a fantasy today. GTA has become a terrible place to live in a very short time period.

brutus1213··on OpenLLaMA: An Open Reproduction of LLaMA
I'm really curious how Meta, DeepMind and OpenAI make the big models work. The biggest A100 you can buy is just 80GB. And I assume the big companies use single precision floating point during training. Are they actually partitioning the big model across multiple GPU instances? If one had the hardware, how many GPUs does the biggest LLAMA take? These are systems issues and I have not read papers or blog posts on how this works. To me, this infra is very non-trivial.
brutus1213··on Geoffrey Hinton leaves Google and warns of danger ahead
To save people the search: https://www.youtube.com/watch?v=7h8jqEs_8kA
brutus1213··on Grand jury votes to indict Donald Trump
Not a lawyer, nor am I American. No horse in this except maybe a citizen of the world. This seems like a weak case to me. For a former president and polarizing figure, you would think the bar on prosecution would be air tight. If he goes to jail for a crime the population (or his support base) considers is trivial or if he does not get convicted, it raises Trump's stock to the stratosphere. Perhaps if he goes to jail, he can't run for office? Still .. seems very dubious based on the info in the news so far.

Here is an analogy I thought of. Say my firm's database gets ransomewared. If I pay the ransom directly, that may be a particular way of accounting. But say, I pay McCohen Computer Associates to fix the problem. Why do I not book it as a computer expense? I am not an accountant but am curious how this works.

Also, booking something in the wrong accounting category, doesn't seem like a very convincing charge. On the other hand, Tax and election violations are very serious.

brutus1213··on The state of Computer Vision and AI: we are far away (2012)
This is somewhat solved. I believe this was a test image for Flamingo (from DeepMind).
brutus1213··on Ask HN: What projects did you build to get better as a programmer?
This is an excellent topic and I was hoping to find a set of great ideas here (really love the http server example). I was thinking of the concept of Katas in martial arts. What would be the Katas for programming I wonder.

Some thoughts:

1) Build your own database. Not a SQL db. Just something that takes records on a cli and enables retrieval. Persist the data. Support better queries/

2) Maze solver.

3) Snake game

4) Helper/Solver for something like Wordle.

5) Chat server (but I like the HTTP server idea. Annoying thing is newer HTTP versions have gotten complicated).

I think there are twenty to fifty of these little things we have all done in our career, and either forgotten them or it felt too natural to register. Wish someone compiles a list of such projects some day. It would bring back joy in some of the coder-at-heart folks lives.

brutus1213··on Ask HN: Does anyone here have an Amiga 1000 they are willing to sell?
a500 was amazing. But why buy one now? Emulators are pretty good.
brutus1213··on Anyone else witnessing a panic inside NLP orgs of big tech companies?
Look .. I am one of the working class here. But I gotta point out. Our standard of living is far superior to someone living in 1970 when it comes to stuff/technology. The stuff is cheaper because of China/globalization. The tech is there because we outsourced production and kept specializing the work force.

I don't think UBI is the solution. Nor is squeezing people more than they are being squeezed. Efficiency and productivity are good things. What is wasteful are things like make-work programs like the DMV or other govt office. That crap needs to be automated away. Hospitals need more funding. Schools are unclear. I think schools would benefit from privatization. I don't think the same of hospitals. Not sure why.

brutus1213··on Ask HN: Has anyone made the transition from app development to ML/AI work?
Just wanted to add: DL Engineer != MLOps engineer. DL Engineers who deal with scientists have a much harder job IMHO as the scientists ask for 10 different things, each taking a lot of effort to implement. The current DL stacks (e.g. Tensorflow/keras) make it very easy to implement the standard stuff. But if you want to go off the path, it is really very hard (I wonder if Pytorch makes that significantly easier if anyone cares to chime in).
brutus1213··on High-value Amazon orders 'switched for cat food', say customers
In Canada, was asked to ship back some exercise supplement.
brutus1213··on My dad died last week and now I am the only male in the family.How do I procede?
I'm sorry for your loss and the difficult circumstances your family finds itself in.

Not sure what to advise you. This is a very challenging set of circumstances you are in.

Two thoughts:

1) You must keep it together as the alternative is unproductive.

2) I believe a defining power of humans is our ability to adapt. You will adapt. So will your family.

Just do the best you can. Work hard to earn but be careful to not self-destruct. Think of an engine that is pushed beyond its limit. You must not damage yourself through too much stress and work.

Don't make false promises such as "Don't worry ... I have it all figured out" .. especially if that is not true. Discuss the reality of the situation and your plan. This could mean no new iPhones or anything beyond the bare necessities of life (not forever but at least for a while).

Remember, you will hopefully earn more money and have an easier job as time goes on. Upskill yourself as you get the opportunity and avoid bad habits.

Best of luck to you.

brutus1213··on Introduction to Data-Centric AI
Thx for sharing the link to your work. Interesting idea. A few thoughts:

1) Congrats on the birth of your child. I imagine you have zero time now but as they get older, you start getting your time back. I went through this and now I can sneak in personal projects while the kids are in their activities, late at night, early am. Be aware your body and stamina declines as you age. I am getting close to mid 40s and I can feel it.

2) The previous poster presented an interesting idea about putting a basic classifier in the loop. The challenge is how do you if the classifier gets it wrong. Confidence scores you get from the logits are extremely flakey. I think one solution is to metric learning methods (contrastive loss instead of cross entropy). I have seen some papers that dance around this but have not seen anything fully baked from a scientific perspective.

3) Your task is an interesting action recognition task. You should seriously consider putting it on kaggle or write a paper on it (and release the dataset). The easy off-the-shelf model you could try on this data for a video classification task like this is possibly X3D. But there are a variety of other methods (I'm a researcher in the field).

brutus1213··on Another Round of Layoffs at Meta
Interesting comment. Curious who you see to be the winning horses, or at least candidates. OpenAI is certainly something special. Not sure who else I'd add to the list. AirBnB has a really strong culture of experimentation and infra for it. Their product is boring though. From your perspective, is it about culture, product, people, something else?
brutus1213··on Another Round of Layoffs at Meta
But that's why I have Peloton and other gym equipment at home.
brutus1213··on Ask HN: How do I get back into the tech industry after 4 years of unemployment?
My take:

1) You have an undergrad degree. The point of the degree is not to teach you knowledge one shot. It teaches you how to learn and problem solve. Your post has a tone of desperation. You need to realize what amazing qualities you bring to any kind of employment situation. Just this alone will make you successful.

2) In your situation, just don't even bother with Leetcode problems. If an employer flunks you for it, so be it. Thank them for the interview practice.

3) You need to apply consistently for jobs. They do not need to be fashionable. Make a website for your local tiny business. Heck .. go to your favorite tiny restaurant that doesn't have a site and ask to make it to them .. pro bono. If you did that for 1 month on your own terms, you will get a resume going.

4) I appreciate the comments posted by people suggesting OSS projects. I think that is too much for you and too long of a path. I recommend making portfolio pages or apps. Remember how you did it in undergrad getting that first job?

5) Stick to boring, mass market tech. Python and Java are great for employment. I would not recommend time with Niche languages unless you want to shoot for the stars right away. Do brushup on git.

6) Don't stress out and don't work crazy hours on anything. Start with 2-4 hours, and build up to 8.

Stay healthy

brutus1213··on Ask HN: Software Devs who lost their jobs and got rehired, where did you end up?
I luckily was not affected (yet) but also keeping an eye on people's tips.

Consulting shops seem to be hiring

Govt. agencies, especially at the local level

Universities are hiring (STEM profs in the US are flush with grant money it seems)

Mom and pop-type local business seem to be growing their small teams. This is the hardest nut to crack as I doubt they use recruiters, LinkedIn, Indeed.

In terms of sectors, most things seem to be negatively affected but some sectors less so. This may also be influenced by your local area.

Best of luck and stay positive!!

brutus1213··on The fraudulent claims made by IBM about Watson and AI (2021)
Sorry .. what is a "Kelley better"? I'm intrigued.
brutus1213··on Ask HN: Employers, why do you want us back in the office?
I was talking to a CTO at a mid-sized startup (200 people or so; private company; I think ex-YC but not 100%); been there for a decade type concern.

Person flatly said no one works in WFH. I am also a manager and responded my team busted their balls the whole time, and delivered results. My personal opinion is my team did 120% of their productivity. I also acknowledge that some people are not built for it and their productivity may fall. I believe a manager can deal with this (either creating the right task-incentive structure, or getting the person off the bus). After my impassioned argument, the person said ..nope .. people just don't work during WFH. This was a social event, just the two of us talking, so I am quite sure this guy was being honest.

I also ran into a different person in a leadership position (just not very high; more like technical leadership; ex-professor; stints at Google). He was convinced the real reason FAANG hired so crazily was because productivity per employee had dropped dramatically. His belief was no one worked during WFH. Again, my impassioned arguments and personal anecdotes did not really move this person.

brutus1213··on Show HN: I trained an AI model on 120M+ songs from iTunes
How did you get the samples for the song? iTunes allows scraping?

Your project is extremely motivational .. how long did it take you? What did you train on? I do DL for work and just play with things like cifar. This is so inspired.

brutus1213··on The Coming Care Crisis as Kids with Autism Grow Up
I've been thinking about this issue a lot. I wish there was a gig worker type role people with autism could play. If they had meaningful employment, it would change their life. Despite some concerns, I think Goodwill does a decent job providing simple jobs/job training to some folks with autism.

On a separate thread, what can tech do to make things better? I don't see this as being venture funded .. it has to be grant based at the start. Wish large, successful folks like YC had a special focus on this sort of tech .. we need more of this rather than TikTok clones.

The spectrum nature of autism makes this so so complicated. We really need a ton of research at understanding this.

brutus1213··on Ask HN: Why do people think the USA did not go to the moon in 1969?
Was telling my little kid that some people still believe the earth is flat. She immediately brought up: - haven't these people used aeroplanes? - we can see how the timezones wrap around - what happens if you go over the edge if the Earth was flat. - we can the Earth is like a marble when we look at it from space (e.g. a satellite)

It truly boggled my mind how the flat earth theory can be plausible in this day and age.

brutus1213··on Ask HN: How to Deal with ChatGPT Anxiety?
I had a teacher in undergrad (over 20 years ago) who said UML and code generation are getting so advanced, 5 years out, you likely won't be coding anymore.

In the early 00s, there was tremendous outsourcing going on. Some people said Western programmer salaries will collapse.

So your anxiety is nothing new. I think of ChatGPT-like tech to be like grammar checker. It will make your job easier.

Also, it is interesting to consider how chip design works these days. My understanding (as a software person) is that there are more people doing test-verification harnesses/QA of sorts than designing the basic blocks. This model may be how things go in software.

brutus1213··on Ask HN: Techwise, how did TikTok get so good?
Here you go: https://arxiv.org/abs/2204.12954 .. not from TikTok but some researchers who reverse engineered some network aspects. Thought the paper is well worth the read.

Systems research people are not valued enough .. my understanding is a bunch (not sure about number) of ex-MSR folks went to ByteDance .. sizable pay bumps.

brutus1213··on Tell HN: Doing a chargeback against Google will get your account banned
What country? What bank???
brutus1213··on The failure of Amazon's Alexa shows Microsoft was right to kill Cortana
Not impressed by the article. The heading raised an instant question .. when did Microsoft kill Cortana? I tried figuring that from the article but it seems to be pointing at tea leaves. The article links to a history of Cortana article but that does not indicate it is dead by any means.

Edit .. I googled and seems to see stories about Microsoft removing Cortana mobile apps on Android and iOS, and some removal of features. I call this focusing. I don't see evidence that the backend infra or Windows-level support is going away.

brutus1213··on Ask HN: Getting into AI?
Parent said they were joking about that example. Not every STEM PhD is math heavy. CS itself is big field with subareas that may or may not require fluency in topics such as optimization (e.g. HCI, Databases). ML/DL is optimization heavy, no? So the point is valid I think.
brutus1213··on Clara.io Shutting Down
Good points. One way to compete with blender could be simplicity. I want to encourage myself and my wife+7 year old to get into 3 D modeling (basic level stuff so we have fun with it in VR/AR or maybe print out models). It all seems a bit inaccessible without some serious effort. When there are too many features, taking some out + not having to do an install is compelling.
brutus1213··on Clara.io Shutting Down
Thank you for being so forthright. I'm curious if you considered lowering the price to where it became stupidly obvious for a user to get. E.g. 1 dollar a month (or year). I recall Whatsapps early pricing model where I'd never pay for a paid messaging app personally, but since it was free the first whole year and 1 dollar a year after, it seemed essentially free.

When I see something that's 10 dollars a month (100 w/ year discount), I think of the higher number as the price (x1.5-x2).

I can see the danger of devaluing the product, not covering costs, etc. but am curious if a price point between free and monthly exists .. ultra-good-value? Use that to bring in some recurring revenue, and grow from that base.

If you have thoughts on this, would appreciate hearing them.

brutus1213··on 83% of Developers Suffer from Burnout, Haystack Analytics Study Finds
I wonder if the economic situation affects burnout. Managers push more to keep their jobs and protect their team. Individual team members push more of themselves to be seen as more productive. You only get productivity pushing yourself to a point .. after that is the redline zone and burnout.

This fear effect is compounded as you hear stories of bullweather employers doing hiring freezes (aka. AWS). I think we have a problem right now in tech .. too many people have been laid off to be absorbed in the same high-tech jobs. People will need to trade downwards, go to grad school, whatever. Not sure if I am missing some bright streaks of light in the cloud.

← PreviousPage 4 of 14Next →