88 karma · joined February 20, 2022
Humans hear car/road noises, along with potential screams from outside or passenger shouts from inside, we sense vibrations, can respond to pedestrian or other driver hand signals, and constantly predict hazards through perceptions. How is the car doing all of this, if not for additional sensors and processing?
You can land a plane by eye, but what happens when there's fog? That's exactly like the situation in cars. LIDAR can provide extra sensory data where the cameras absolutely fail, just like our own eyes.
Knowing there's a solution to this, we are just to accept the car will fail where humans will? That's progress? Why wouldn't you want that extra data for such a small relative cost? LIDAR was already used on cars as a safety-only front collision avoidance system (that's how cheap it is to install).
In a properly designed system, adding data which is useful and cannot otherwise be inferred makes complete sense.
Given these cars are supposed to be so good that they will be working autonomously for you and pay themselves off in a year or two, the idea that LIDAR etc. is unnecessary and too expensive and will be lead to actively worse performance, is just insane logic for an "engineering" discussion.
Consider how many car manufacturers are backtracking, the fact that China will now win the EV race, utility energy prices have skyrocketed, and the damage done not only to the brand but people's view towards EVs in general.
Forcing this play happened too soon.
This is compounded by the system mistakes likely being hard-errors. A computer hard-error vs. human lapse of judgement is potentially the difference between the vehicle slowly crushing a small child as they scream and beg for help vs. a human stopping as soon as they felt/heard something. Context matters.
The compared error-rates must consider if it could have been avoided or mitigated, the near misses, the human vs. computer type of error, and how hard-errors may lead to horrifying scenarios.
The idea that mostly-automating the system because it's statistically better than humans, but requiring human-assistance to monitor and respond in these exact situations, was flawed logic to begin with. Comparisons of statistics should be made like-for-like, given these are scenarios we can easily control.
For example, a robotic taxis should at least be compared to professional drivers on similar routes, roads, vehicles, and times of day. Not just comparing "all drivers in all vehicles in all scenarios over time" with private company data that cherry-picks "automated driving" miles on highways etc. (where existing assistance systems could already achieve near-perfect results).
Companies testing autonomy on the public should be forced to upload all crash data to investigators as part of their licensing. The vehicles already have extremely detailed sensor and video data to operate. The fact that we have no verified data to compare to existing human statistics is damning. It's a farce.
For example, Apple and Google should provide tools for the parents to set up a device appropriately for a child, much like the shop should not sell alcohol to underage customers. Similarly, content producers should specifically need to label content targeted for children or specially 18+, like the producer of alcohol must warn customers on the label and inform the retailers.
Parents and caretakers need information to make an informed decisions before being able to consume the media themselves. They also need some granular tools on the device to avoid banning them entirely. The burden is shared between creator, distributor and consumer.
We already had laws for this and it makes sense for some type of access control to the open internet. The shocking part is the requirement for everyone to verify ID to multiple public and private institutions, more than once per.
An analogy for the UK now would be needing ID to enter the supermarket (access the internet), ID to look at anything aimed at adults and potentially harmful such as alcohol, chemicals, sugary food, medicine etc. (know "potentially harmful" subjects exist), ID to look at anything lawfully 18+ such as alcohol and cigarettes (view the content), then ID again to make the 18+ purchase from an account needing ID to open.
Perhaps some people can sense this stuff subconsciously. Relationships should build naturally.
Anyway, what are you paid for? Guessing a programmer, you just sit in a chair all day and press buttons on a magical box. As your customer, why am I having to explain what product I want and what my requirements are? Why don't you have all my answers immediately? How dare you suggest a different specialism? You made a mistake?!?
When a businesses chooses to drop support for a product entirely (hardware or spares no longer produced, and software no longer updated), they've presumably already made the business decision to drop the product for sale. If the product were still in demand and the existing devices still function, dropping product support could effectively render the devices useless or destined for landfill.
This often happens when: online services are dropped, devices cannot be repaired, or worse, software cannot be simply updated for security and compatibility reasons etc.
If manufacturers want existing users with working technology to upgrade, they should design compelling improved products, not force a load of e-waste and bricked devices. There's little reason for a manufacturer to quickly drop support without following this model of open sourcing, unless they know they are forcing existing customers to an unnecessary upgrade.
Manufacturers should support their products, innovate, or let them go over time. The "market" (consumer choice) should dictate when a product is obsolete. We own our products and should have the rights to maintain them. They should be paying us and taxed for damaging the environment for dropping support early and/or without open-sourcing.
It's harrowing how people prefer donating resources over exerting mental effort to bridge simple psychological boundaries in understanding the different needs of others, especially for disabilities (which nobody chooses to have). I often wonder if the root of this is the individual fear it could happen to us. By exercising empathy, we are reminded that ourselves and our families are vulnerable to disability at any time--from birth to life events this second (injury, illness, luck), existence is vulnerability.
Our intrinsic fears combined with societies lacking safety nets and breathing space has created a positive feedback loop for hyper-individualistic living. Our own bubbles. I try to do the opposite, but it's not easy.
The main task of existing tools is rule-based checks and flagging errors for attention (like a compiler), because there is simply too much for a human to think about. The rules are based on physics and manufacturing constraints--precise known quantities--leading to output accuracy which can be verified up to 100%. The output is a known-functioning solution and/or simulation (unless the tool is flawed).
Most of these design tools include auto-design (chips)/auto-routing (PCBs) features, but they are notoriously poor due to being too heavily rule-based. Similar to the Photoshop "Content Aware Fill" feature (released 15 years ago!), where the algorithm tries to fill in a selection by guessing values based on the pixels surrounding it. It can work exceptionally well, until it doesn't, due to lacking correct context, at which point the work needs to be done manually (by someone knowledgeable).
"Hallucinogenic" or diffusion-based AI (LLM) algorithms do not readily learn or repeat procedures with high accuracy, but instead look at the problem holistically, much like a human; weights of neural nets almost light up with possible solutions. Any rules are loose, context-based, interconnected, often invisible, and all based on experience.
LLM tools as features on the design-side could be very promising, as existing rule-based algorithms could be integrated in the design-loop feedback to ground them in reality and reiterate the context. Combined with the precise rule-based checking and excellent quality training data, it provides a very promising path, and more so than tasks in most fields as the final output can still be rule-checked with existing algorithms.
In the near-future I expect basic designs can be created with minimal knowledge. EEs and electrical designer "experts" will only be needed to design and manufacture the tools, to verify designs, and to implement complex/critical projects.
In a sane world, this knowledge-barrier drop should encourage and grow the entire field, as worldwide costs for new systems and upgrades decreases. It has the potential to boost global standards of living. We shouldn't have to be worrying about losing jobs, nor weighing up extortionately priced tools vs. selling our data.
I also find [flagged] content appears on HN aggregate sites, and it shows the inherent censorship.
It's unfortunate that considering supposedly only "seasoned" participators of HN can downvote, it runs deeper than a surface issue. You will be downvoted without comments replying with the counterarguments.
Scientific viewpoints often make way to hope and cope engineering here. This will only work as long as the people involved are insulated from the direct effects of their actions.
It's a shame. HN felt like one of the last bastions of the old internet where techies came to discuss tech, science, and occasionally important worldwide news, in a technical and objective way.
Company blog/announcement -> developer blog -> HN front page -> Twitter -> small tech blogs -> YT niche videos -> Reddit -> PR newswire -> bigger (paid content) blogs -> YT general/news videos -> Facebook -> local news -> national news.
Meanwhile, newer developers who think they're early on groundbreaking tech news are spreading this to their friends via Discord and Whatsapp etc.
Yet all of this media is controlled, just to generate buzz and boost stock prices.
Partly is seems to be how quickly an article has comments leaning to one side or another. Once a few of these comments get off the ground, it's hard for voting on HN to reflect the discussion these days.
It does appear that more users from Reddit etc. are not just using HN these days, but commenting. The quality of posts and comments has definitely decreased, in line with the quality of content on blog posts decreasing.
Technological advancements in modern times appear to only benefit very few numbers of individuals. The 1960s excitement was a time when we imagined it would benefit everyone more from it--not necessarily equally, but at least life not getting worse! Now we just see ourselves being replaced and struggling more, with the "lucky" ones in precarious employment.
Whilst wealth inequality is already creeping towards record levels in the West, with potential wars are on the horizon, civil war being encouraged, and tech stocks appearing to be in a bubble alongside P/E ratios of corporate stocks at high levels like we saw before other financial disasters, it's hard for anyone to be excited about the near future.
A lot of our best science and technology emerged from bad intentions--war, accidents, control, private power struggles. A lot of our worst science and technology emerged from good intentions--charity, humanity, heroism, public works. It's a strange world. We have to contend with edge cases, the butterfly effect, and dynamic chaotic systems etc. These are some reasons for the potential "road to hell paved paved with good intentions".
The state of AI is unfortunately tragic to me. I've always loved, learned, used, designed, and then built technology. My life has largely revolved around it, and I'm witnessing another technological revolution. Yet the more I've learned about how humans use and control technology, the more I realized the truth:
Technology =/= progress.
We should all think extremely carefully about what we're contributing to. Progress is inevitable, but as designers, builders and preachers of the latest technologies, we have a small amount of power in society to help ensure resources cannot be secretly utilized against us. This small power is in our choice to use current offerings or encouraging others to, and whether we work for those who with to control us.