HNHacker News
TopNewBestAskShowJobs

einrealist

1,436 karma · joined December 16, 2013

Former CTO at a startup. Now a senior consultant at Evil Corp, again.
submissionscomments
einrealist··on Duolingo CEO tries to walk back AI-first comments, fails
Good times for ex-contractors of Duolingo to have AI help them build a honest competitor app. Just watched a video about Sears today and it reminded me how fast top dogs can fail.
einrealist··on Positional preferences, order effects, prompt sensitivity undermine AI judgments
"and keep a human in the loop whenever single-digit accuracy points matter"

So we are supposed to give up on accuracy now? At least with humans (assuming good actors) I can assume an effort for accuracy and correctness. And I can build trust based on some resume and on former interaction.

With LLMs, this is more like a coin-flip with each prompt. And since the models are updated constantly, its hard to build some sort of resume. In the meantime, people - especially casual users - might just trust outputs, because its convenient. A single digit error is harder to find. The costs of validating outputs increases with increased accuracy of LLMs. And casual users tend to skip validation because "its a computer and computers are correct".

I fear an overall decrease in quality wherever LLMs are included. And any productivity gains are eaten by that.

einrealist··on Claude 4
And how many times did this scenario not occur?

Guess, people want to focus on this particular scenario. Does it confirm biases? How strong is the influence of Science Fiction in this urge to discuss this scenario and infer some sort of intelligence?

einrealist··on What Sam Altman told OpenAI about the device he's making with Jony Ive
Can the device work offline or does it need constant connection to a data center? Already sounds like a security and privacy nightmare.
einrealist··on Watching AI drive Microsoft employees insane
This is one good example of the Sunk Cost Fallacy: generative AI has cost so much money, acknowledging its shortcomings is now becoming more and more impossible.

This AI bubble is far worse than the Blockchain hype.

Its not yet clear whether productivity gains are real and whether the gains are eaten by a decline in overall quality.

einrealist··on OpenAI is building a social network?
This is Altman increasing the mass of the investment black hole that OpenAI is.
einrealist··on Recent AI model progress feels mostly like bullshit
That "good enough" is the problem. It requires context. And AI companies are selling us that "good enough" with questionable proof. And they are selling grandiose visions to investors, but move the goal post again and again.

A lot of companies made Copilot available to their workforce. I doubt that the majority of users understand what a statistical model means. The casual, technically inexperienced user just assumes that a computer answer is always right.

einrealist··on Recent AI model progress feels mostly like bullshit
LeCun criticized LLM technology recently in a presentation: https://www.youtube.com/watch?v=ETZfkkv6V7Y

The accuracy problem won't just go away. Increasing accuracy is only getting more expensive. This sets the limits for useful applications. And casual users might not even care and use LLMs anyway, without reasonable result verification. I fear a future where overall quality is reduced. Not sure how many people / companies would accept that. And AI companies are getting too big to fail. Apparently, the US administration does not seem to care when they use LLMs to define tariff policy....

einrealist··on Why AI is still dumb and not scary at all (pt. 1)
I asked myself this question today: Given the AI hype, I wonder whether there will be employers who force their employees to use generative AI. Otherwise employees will be deemed inefficient and be fired.

Employers might believe that not using AI puts them and their investments at a disadvantage, while ignoring the overall quality issues.

einrealist··on The Government Knows AGI Is Coming
Boring how there was no questioning of the goal post that is moved constantly. This (generative) AI industry suffers from an absurd level of sunken cost fallacy.

Unless there are more fundamental changes in the technology, I only see us in the biggest bubble of the modern world.

einrealist··on Gödel's theorem debunks the most important AI myth – Roger Penrose [video]
That's how I understood Penrose's reasoning too. He differentiated between the computer and whatever is going on in our brain. Computers are just "powerful" enough to encode something that mimics intelligence on the surface (the interviewer tried to pin him on that "something new"), but is still the result of traditional computation, without the involvement of consciousness (his requirement for intelligence).
einrealist··on The A.I. Monarchy
You are right. I oversimplified my statement to make the point that the resource requirements of modern economies are more complex. So conflicts happened over less complex resource requirements back then. Also resources are not just the source materials, but anything within a supply chain, including intermediate products.
einrealist··on The A.I. Monarchy
And those small kingdoms did not care about oil, minerals back then - things that might not exist in your vicinity, but are required to run a modern economy today.

This entire concept of city-states exists on a premise of peaceful coexistence and cooperation. Given our history, this is pure fiction. (And then these people shit on Communism.)

einrealist··on The A.I. Monarchy
This network of (almost) city-states existed before: https://en.wikipedia.org/wiki/Kleinstaaterei

It was a time of wars and economic turmoil.

einrealist··on GPT-4.5
And remember, with increasing accuracy, the cost of validation goes up (not even linear).

We expect computers to be right. Its a trust problem. Average users will simply trust the results of LLMs and move on without proper validation. And the way the LLMs are trained to mimic human interaction is not helping either. This will reduce overall quality in society.

Its a different thing to work with another human, because there is intention. A human wants to be correct or to mislead me. I am considering this without even thinking about it.

And I don't expect expert models to improve things, unless the problem space is really simple (like checking eggs for anomalies).

einrealist··on Modern-Day Oracles or Bullshit Machines
You are right. But these are different different types of motivations of the same thing. And there is always context for these motivations.

Its a different thing to sell Trump that LLMs should take over crucial decisions within a government than just using it for some prototyping, code completion at work or to create cat pictures at home.

Take Copilot for example. It was rolled out in different companies, I worked with. Aside of warnings and maybe trainings, I doubt the companies are really able to measure the impact that has. Students are already using the technology to do homework. Schools and universities are sending mixed signals about the results. And then those students enter the workforce with Copilot enabled by default.

At least with companies, its the "free market" that will regulate (unless some company is too big to fail...)

einrealist··on Modern-Day Oracles or Bullshit Machines? How to thrive in a ChatGPT world
This website is so important!

Now ask yourself why AI companies don't want to be regulated or scrutinized.

So many companies (users and providers) jump on the AI hype train because of FOMO. The end result might be just as destructive as this mythical "AGI".

Edit: I am not saying to not use the technology. I am just on the side of caution and constant validation. The technology has to serve society. But I fear this hype (and ideology) has it the other way around. Musk isn't destroying the US government for no reason...

einrealist··on What's happening inside the NIH and NSF
Fraud can be proven. But its a process and its not as easy than just claiming something to be fraud. We (hopefully still) have a justice system, with due process for the same reason.

That $9B is just a number. Whether it is reasonable or not can only be determined by looking into the details:

- What accounts as overhead?

- What were decisions that lead to that overhead?

- Were there alternatives that would have costs less?

- Why and how were those decisions made?

- Can we learn for future decisions?

- Was there actual fraud?

- When there was fraud, why wasn't it referred to DOJ to be investigated properly?

"Administrative overhead" is not bad by itself. Outside government, its simply called "Operating Expense" and "Cost of Revenue" (not a concern of government luckily). I am certain, if you look into SpaceX' or Tesla's expenses, you would find fraud too.

Because Musk and his brainwashed followers can claim stuff on Twitter doesn't make it true. And it certainly MUST NOT be basis to destroy Democracy and democratic processes.

einrealist··on What's happening inside the NIH and NSF
https://www.nytimes.com/2025/02/06/business/trump-politico-u...
einrealist··on What's happening inside the NIH and NSF
There is ZERO context in those Tweets. Only showing a line item, with a dense description is NOT evidence for fraud. Its like if I were looking at your bank transactions and assuming the worst.

Don't get me wrong. There is certainly fraud and overpayment happening in government operations. But just looking at receipts is not the smoking gun. If anyone suspects fraud, there are a processes. Inspectors Generals are (were) one way to have those payments investigated, DOJ and FBI would be another step. The IGs were fired last week. I wonder why...

Also, this "Ian Miles Cheong" guy is literally a Kremlin operative - really someone worth blocking. So that's important to know about his motivations to sow doubt in US Democracy.

einrealist··on What's happening inside the NIH and NSF
- He claimed the US funded bio-weapons when they found payments for gain of function research.

- Calling payments to non-profit organizations fraudulent on a whim.

- The sweeping condemnation of what USAID was doing.

- His call for a blanket drop of regulations.

Either he knows better or he is totally lost in his sauce. Hard to say what's worse for where he is right now.

I believe, he does not care. He only cares about his conception of the world and how AI and Mars are more important than those tiny tiny human problems. Society has to serve him and his god complex. He was told to find his subsidies and tax cuts by himself. That's what he is doing.

einrealist··on What's happening inside the NIH and NSF
The process is the problem. There is no oversight and accountability for Musk and his "DOGE". That's pure poison to Democracy and to a functioning society.

Musk is neither competent nor efficient. He looks at line items and makes stuff up. He destroys a hundred useful things to destroy a bad one. Details don't matter to him. Its the same con man mentality that feeds off the works of his workforce. People who think he is a genius are gullible.

einrealist··on Germany: Number of unemployed highest in 10 years
Storage capacity has to be built along with solar and wind parks. Its already happening. But more investments are required ("more effective" than nuclear). And the electric grid must be decentralized further to accommodate decentralized storage (from your wall-box at home to industrial storage facilities). And it must not stop at borders.

Regenerative energy already surpassed all other means of energy production in 2024: https://www.bdew.de/media/documents/2024_12_18_Die_Energieve...

Battery prices are declining for some time now: https://www.bloomberg.com/news/articles/2024-12-10/electric-...

And even larger storage capabilities are built: https://standort-sachsen.de/de/aktuelles/news/detail/n1474-g...

(I am actually working for a grid operator)

einrealist··on Germany: Number of unemployed highest in 10 years
I am glad that we are afraid. The real costs (no subsidies, fuel, waste and dismantling fully paid by owner) of nuclear are just too high. Hinkley Point in the UK is a warning. Investments are far more effective in solar, wind and storage capabilities. Market prices reflect that very well already.
einrealist··on Nvidia, ASML Plunge as DeepSeek Triggers Tech Stock Selloff
Nice to see a pullback within this AI bubble. Wake me when I don't have to double-check the results of generative AI, even with DeepSeek.
einrealist··on Trae: An AI-powered IDE by ByteDance
And a world, where you cannot be sure who has access to your source code (or even to your systems).
einrealist··on Stargate Project: SoftBank, OpenAI, Oracle, MGX to build data centers
That‘s why the tech oligarchs told Trump that Canada is required. Cheap hydroelectric power…
einrealist··on United States Digital Service renamed to DOGE
No. If the content is thin, make a large headline.

(Worked well throughout Trump's campaign. They even printed it on red caps.)

einrealist··on TikTok goes dark in the US
The law and also the SCOTUS ruling don't even touch editorial discretion. TikTok is free to do whatever as long user data is safe from being accessed by an adversary like the Chinese. Its TikTok who made it about editorial discretion and free speech. They failed with that argument. And I agree with that.
einrealist··on TikTok goes dark in the US
The US government claims otherwise. And even if data is not flowing to China directly, the threat is there from a legal point of view. Data is leaving US jurisdiction anyway. And China could force ByteDance to hand over data that is stored in Malaysia and Singapore.
← PreviousPage 5 of 19Next →