n_ary

The best thing about EBs is, they do not hum and vibrate so much making it feel like the bus is about to transform into some Decepticon bot any minute now.

It took many years for online tracking, iframes, sticky cookies and cambridge analytics before things like GDPR came into existence. We have to similarly wait a few years before similar major leaks happen through LLM pipelines/integrations. Sadly, that is the reality we live with.

Is this the beginning of the apps for everything era and finally the SaaS for your LLM begins? Initially we had internet but value came when instead of installed apps, webapps arrived to become SaaS. Now if LLMs can use specific remote MCP which is another SaaS for your LLM, the remote MCP powered service can charge a subscription to do wonderful things and voila! Let the new golden age of SaaS for LLMs begin and the old fad(replace job XYZ with AI) die already.

> Wonder what the kids of 2060 will be throwing away as a result of our life-experiences.

EOL devices(tablets, phones, macbooks, thinkpads, hobby electronics boards, home lab equipments, hdd and ssd full of archive data, swag from conferences, outdated books on product and programming, smart watches etc).

It depends… There are really good people and there are bad people. While there is room to improve, I have decent number of colleagues who praises the system and decent number who have points that they deserve more respect. That being said, there is a secret saying, if you hate foreigners then you work in immigration to make their life hard and derive pleasure. Other than that, the rules change too much than necessary, because immigrants are the powerless scapegoats for any political and social failure of the nation, so the case workers are always playing catchup to the new changes and commit mistakes both unintentionally and intentionally(to avoid legal issues).

Here is the example of why LLMs need to replace some douchbag SWE. But then again, unless there is a nutcase product owner involved, update check every 5min and downloading automatically is pure symptom of bad engineering practices and fear driven development(if anything breaks update is 5min away and we can laugh at the user when their app is updated and they are baffled).

Good thing, this was not shopify/Duolingo/Msft, else the news would be, how AI saved us $8k by fixing a dangerous code and why AI will improve software quality.

I find the misbehavior of indie/boutique MacOS apps always insisting on starting at login very irritating. Unless the app needs to run some heavy background preparation steps before becoming usable, there is literally no sense it starting at login. Also when dormant, check for update(once every 24h), and nag the user if they want to update, but please do not auto download! A lot of non-tech folks use 128/256GB versions of macbook with trillions of photos already clogging their device, an app downloading new updates to add to the pain unless the user asks to do so is outright malice.

As another fellow European, I also agree that this is putting nazis and dangerous people behind bars but recently more generic people had gone behind bars or legal hurdles compared to actual nazis amd racists because due to heavy shift to right, racism and nazi supporters now pretend to be doing political campaigns to cover their crimes and walking free. Not a good sign but I am optimistic.

Unless this is a sarcasm, you have proven that either you are 150% disconnected from reality or is discounting the entire immigration hurdle and access to every basic necessity(bank, housing, healthcare, work permit, qualification recognition etc) requiring a valid ID.

Hmm... But SN himself is not writing code or involved in day to day development. He can claim anything and his incentives are 100% aligned to claiming AI bs.

In similar development, as much as 98% AI generated code wastes my time because I need to delete or press escape to stop the auto completion because a) it is spewing code from official doc verbatim, b) risks injecting licensed code into the product and getting my employer on legal fire, c) it is wrong and exactly not what I want.

> We will likely encounter world order disrupting changes in the coming decades when computer get another 1000 times faster and powerful in the coming 10 years.

Sadly for several decades, as computers got faster, only tracking and ads got more pervasive. Also as hardware became stronger, computer games, media quality improved. We also got planet burning blockchain and crypto mining for a while.

In terms of intelligence, we started with clippy, Eliza, stayed with Cortana and Siri, now have LLMs. We will move forward with LLMs for a while until something more impressive comes out and mirrors more humane nature.

Also GenZ is literally second worst off generation after the silent generation, growing up in unprecedented prosperity to be only be stunted by global pandemic following by crazy oligarchs trying to bring back techno-feudal age while their parents prepare to sell of the last glimmer of inheritance to finance retirement and healthcare. I honestly do not blame them.

> “why they cannot get what they want done using AI.”

Or in other words the CEO had proven entirely out of touch with the actual work and how far away he is removed from entire company operation.

It is hard to tell at this point about the AI/LLM. But the powerful hardware backing it has numerous potential in other research and innovations which are not yet clear to us. Just like the fiber and cellular network made Meta or Airbnb obscenely rich, the hardware may as well facilitate other forms of tech, may be quantum computing or heck even blockchain may be back(surprisingly these are suddenly very legal and less regulated this year!) or who knows what might be new.

While LLM trend is already going to the dumpster(notice how we no longer receive 1b/4b/8b models from Meta since LLama4 and similar competitors but only from china?), I firmly believe that commodity hardware will improve in this race to allow running LLMs(or their next iteration) on regular devices and become as ubiquitous as Siri/Cortana.

True. But like the dot come bust, ultimate winner may not be LLM or even AI but something else entirely using the left over hardware and infrastructure which was intended for AI.

You are indeed correct according to my opinion. Salesforce went too deep into BlockChain and BigData and probably never recovered from the sunk costs. But to stay relevant they again need to risk another bet but also need to conserve(death by a thousand cuts and all) so they layoff while jumping on AI bandwagon. And how fortunate it is that LLM sellers tout productivity gain(zero backing data but hey) as a benefit, so it also falsely support their failure as a success.

Uber probably has legal framework of operation, ability to quickly obtain permits for autonomous testing, local political connections, special ability to gather roads, traffic, mapping data and overall support with rapid deployment. For waymo to get into a new location/country, establish their presence, navigate the political and regulatory landscape etc will take very long and many difficult hurdles.

I think, we need to step back a little in these discussions. We need to ask, what productivity gains are we hoping to find?

In any knowledge or specialised work, operating a tool faster does not give great results, rather raises the risk of error and quality decline.

Did duolingo once face existential threat because they failed to produce specific feature sometime? Did one of their feature suffer and cause user loss because it took more time for an engineer to write the actual code?

Additionally, beyond formatting and obvious logical errors, every new code should in theory need some human review, which means more automated code means longer review. Assuming code is now produce at 2x, it also balances out that review will now take 2x. Additionally review is much more mentally taxing than putting out one’s thought into code, so risk of bugs and security holes also increases in the long term.

While Software devs cost money, the job involves thinking, and that means it can’t be compared to factory work where someone is standing at next step to simply drill a screw in spotX and the next person will simply put on a cover. Despite years of effort, the attempt to make the process mirror factory floor(did anyone notice the open floor parallel to factory floors?) it failed.

While many hate to see it this way, just like a surgeon will take his time to perform a brain/heart/tumor surgery, a SWE will do thinking, planning, coding and reviews. Giving a surgeon an autonomous bot that can spread the incision area faster or perform the incision faster does not mean productivity gain, it just means the doctor still needs to plan where to make the incision, how much to spread, what to chop off, what to avoid.

You are misreading. The declaration effectively does NOT quantify or prove any productivity gain. It simply alludes to a possibility by forcing to adapt a new tool which is frequently sold as productivity gain. There is no real consensus or mass scale data except businesses with executives heavily invested in the AI adjacent stocks or prefer to be called visionaries like how SAP or Adobe got lucky by moving into cloud and subscription model purely by accidentally following random hype.

Also in most software dev task, writing code is rarely if ever the bottleneck, it is the thinking, endless scrum bullcrap, management alignment, architects being too obtuse to not sign off on simpler solutions to protect their prestige, not filling empty roles, pushing teams to extreme and then suddenly deciding to throw away and start something brand new while asking why can’t the thrown away solution be adapted for the new thing or simply consulting/contracting firms intentionally prolonging the project to milk maximum billable hours.

I have actually read the entire screenshot in that tweet and to me it read 50% as non technical executive pitched by some AI vendor about how cool their tools are and 50% of clueless people finding non existent nails for their new hammer.

AI has benefits, it may give productivity gains, specially if you are sweat shops churning out customization of same ERP or wordpress theme. Just repeating productivity gain, coding landscape will change, life will be different 500 times daily because the seller of LLM(calling them the AI wholesale is deceptive, they are subset) wants you to do that, does not magically make those true. The whole trend is sickening and it needs to die.

Replacing SWE is fine, but if SWE can be replaced, the justification can be used to replace crucial things like doctors, lawyers, auditors and lots of other professionals which will kill us all.

In another news, I intend to uninstall duolingo effective immediately and actually in an effort to support human-first businesses and enroll into a real language course class where I get to talk to real human beings with same goal and have actual exams and people trained in teaching particular language.

For me, Duolingo(uber for Anki flash cards of preliminary words for a 3 day tourist in a new country) was always an odd product. It is very popular among people because they can immediately learn how to say hi/hello, thanks, please in new language but after that, it is akin to learning to swim by reading different tips and tricks, without actually ever touching water or doing the act.

It is not the IM apps, SE region which I suspect the author is referencing predominately uses Whatsapp.

The value in SE is mostly B2C, instead of marketplace feature, most local tiny(or even big ones willing to evade tax by not having any physical presence) businesses will open a small business or general page and publish their wares as posts. Lives will be used to demo products or services now and then. People follow these pages and flock over to buy things.

In a sense, Facebook and Whatsapp are like Amazon/Aliexpress of SE Asia. I was there for 5 months visiting a friend(and recovering from burnout), and number of people using such pages to sell anything from basic clothing to food to services are HUGE! It is literally a huge business hub for people to discover and make online purchases. In summary, Facebook pages are the e-commerce front(due to lack of shopify/amazon and similar operators who can handle logistics and payments) for individual businesses.

There were many journalist reports about this phenomenon several years back, but I am too sleepy and tired to link those.

Or a signal that, someone did not think about the bus factor and future of the project when most of the teams jumped ship.

Actually, here in my corner of EU, only the prominent big tech backed well documented and battle tested tools are most marketable skills. So, React, 50 new jobs, but you worked with Svelte/Solidjs, what is that? Java/PHP/Python/Ruby/JS, adequate jobs. Go/Rust/Zig/Crystal/Nim, what are these? While Go has some popularity in recent years and I can spot Rust once in a blue moon. Anything involving requiring near metal work is always C/C++.

Availability of documentation and tooling, widespread adaptation and access to already-trained-at-someone-else's-dime possibility is deemed safe for hiring decision. Sometimes, the narrow tech is spotted in the wild, but it was mostly some senior/staff engineer wanted to experiment something which became part of production because management saw no issue, will sometimes open some doors for practitioners of those stack but the probability is akin to getting hit by lightning strike.

I also find the intellisense aspect of it good though the price is still too high when my local IDE can do 1/10th of that for a long time.

Additionally, LLMs are sort of using old day google mastery to find the right result quickly to save a huge waste of wading through junk and SEO spam, which translates to productivity but then we are again balanced out because this gained productivity was lost once SEO spam took off a decade back. I am indifferent about this gain again, as anything with mass adaptation tends to devolve in garbage behavior eventually, slowly the gains will again be eaten up.

Same could be said about all the hype and trend that died, blockchain(ignore cryptocurrency part), IoT(not sure what happened), bigData(foundation of current AI regardless of whag anyone says) , app for everything(we indeed have more apps now and everything is junk) were also considered the new water/air/electricity/revolution/disruption.

We in aggregate seem to have developed a collective amnesia due to how fast these trends move and how much is burned in keeping the hype machine going to keep us on the edge. We also need to stop calling LLMs different just like every kid wants to claim mark zuck was diff or bill gates was diff so dropping out like them would make these kids owner of next infinite riches.

After a long decade of fast moving “this will truly revolutionize everything” speech every so often, we need to keep some skepticism. Additionally, the AI bubble is more devastating than the previous as previous money was being spread into multiple hypes from which some emerged silent victors of current trends but now everything is consolidated into one thing, all eggs in one basket. If the eggs break, a large population and industry will metaphorically starve and suffer.

I think, such sayings are more and more being proven wrong. Buy the rumor, sell the news, or buy the dip and hodl long term etc have severely backfired in recent months and many individuals(I personally know) lost huge.

Until we return to ZIRP somehow, there is just too much money in AI to keep the hype going as long as possible to avoid the investor backlash or cause mass layoffs(cutting costs and recovering because we spent too much on AI but can’t get refunds so layoff people instead), as money seem to only chase hype and due to not enough free and openly available money for other topics, all flows to only thing that everyone chases.

When AI becomes equally expensive as a human, layoff humans to continue finance AI because, it is easy to fire and rehire but expensive(more precisely acknowledging failure and bad decision making to hurt leadership ego) to remove and become independent from complex integrations.

Edit: I am just sharing how our CTO responds to the massive push of AI into everything, because integration of a non deterministic system has massive cost and eventually once the thing is made deterministic, the additional steps add expenses which finally makes the entire solution too expensive compared to the benefit. This is not my opinion, just sharing how typical leaderships hope to tackle the expense issue.

Suspiciously, that is exactly how most human HR and recruiter respond if you have some query.

Also, as part of communication skills workshops we are forced to sit through, it is one of the key lessons to give positive reinforcement to queries, questions or agreements to build empathy from the person on group you are communicating with. Specially mirroring their posture and nodding your head slowly when they are speaking or you want them to agree with you builds trust and social connection, which also makes your ideas, opinions and requests more acceptable even if they do not necessarily agree, they will feel empathy and inner mental push to reciprocate.

Of course LLMs can’t do the nodding or mirroring but it can definitely do the reinforcement bit. Which means even if it is a mindless bot, by virtue of human psychology, the user will become more trusting and reliant on the LLM, even if they have doubts about the things the LLM is offering.

The article reads… interesting. A business depends on its demand, competition, scale, operation costs and numerous other factors. If the product is a tech, a technical founder can have immense impact on future success but marketing, customer acquisition, sales, networking, fund raising and other huge headaches are often burdens carried by non-technical counter part.

Of course, the person building something from a vague idea to life wields magical power, but the other aspects of business are more so important.

As an engineer myself, I can see 500 ways I can improve things at my employer’s business, but while I have a technical vision, the non-technical founders of the business has broader picture and what I may wish to spend time optimizing might just be a pocket change issue not worth solving.

We technical folks undermine the mountain of difficulties of marketing and sales as well as interacting with people to convince them to part with their money. Of course, I have seen a fair share of over confident wanna-be founder who would just be next trillionaire if someone would build the 10000th Airbnb clone but for pets so they could sell it and give 5% equity to the builder because their idea was the main thing, but in broad sense of things, tech is just a means to an end for a successful business and simply building is not the 100% of execution step.

Honestly, I am surprised by your opinion on this matter(something also echoed a few times in other comments too). Lets switch the context for a bit… human drivers kill few thousand people, so why make so much regulations for self driving cars… why not kick out pilots entirely, autopilot can do smooth(though damaging to tires) landing/takeoffs, how about we layoff all govt workers and regulatory auditors, LLMs are better at recall and most of those paper pushers do subpar work anyways…

My analogies may sound apples to gorillas comparison but the point of automation is that they perform 100x better than human with highest safety. Just because I can DUI and get a fine does not mean a self driving car should drive without fully operational sensors, both bear same risk of killing people but one has higher regulatory restrictions.

After reading this article, this gives me the "1938 and onward" vibes of particular nation in EU. While the different sources and commentators try to sound official and legal, there are just odd parallels to the particular years I mention. History repeats itself I suppose, just against different groups at different times.