supersour

Thank you so much for this! You have helped me considerably.

Maybe a Show HN? I would be quite interested in seeing the results of this project

100%. I am really getting tired of the narrative that code before LLMs was optimally performant, perfectly architected, completely understood, and bug free...

Does the author not have the experience of working in a legacy codebase that nobody really "understood"? Something sufficiently complex where even the senior SW devs needed to scope out project work and research the codebase for dependencies or potential issues?

I fail to remember a time at LARGE_CORP where even the most experienced developers were able to scope out or design a feature without studying the existing documentation, timing diagrams, etc....

Reminds me very much of this, which I have printed out and placed on my wall at home:

https://www.gurdjieff.org/lannes2.htm

100% agree. I understand the argument around the "fun" of programming and learning. But my joy from coding (especially when working 9-5, on a team of 5-500 people) always came from the value creation for other people and for society.

AI has simply been a tool to accelerate that and I've genuinely never been happier about the value I provide to society. I find it hard to take the "respect of the craft" argument at face value - even if I enjoy knitting by hand, I can still support the existence of mechanized looms and machinery that provide affordable clothing in the millions.

My late grandmother knit 1-2 pairs of socks per day and donated them to the homeless center - absolutely amazing thing to, and it kept her motivated and sharp up to 104 years of age.

However, if her ultimate goal is providing socks to the homeless, would it not be better for her to work a library job for $15/hour, then take the $100+ a day and purchase then donate 30 pairs of socks? Even if the socks are a lower quality and wear out twice as fast, she could help 15x the people thanks to the mechanization and efficiency provided by the devision of labour.

I see this divide more and more and truthfully I wish I could understand both sides. I am encountering trouble in the startup I'm working on for exactly this reason: I feel laser focused on the attempt to create value for others, while my mech eng cofounder has this "respect of the craft". From from my perspective this has only caused delays in generating revenue.

Any help in understanding those who "love to make" (as opposed to those who "love to deliver") would be extremely useful.

Yes, power efficiency may be the largest benefit of these chips actually - especially if the projections are true that the US and other countries simply aren't able to ramp up power generation to meet forecasted datacenter demand.

There's a bit of a ticking time bomb there, something that a Taalas-like architecture can clearly resolve.

Yes this is true to some extent. I've been using LLMs to run some computer vision tests, and I've certainly noticed myself running into the trap of "just one more AI experiment", or "just one more change" while neglecting to actually properly integrate the learnings into my mental model.

However, without an agent running its own experiments on a cloud GPU, would I realistically have invested my limited work hours and tried evaluating 10 different models, each with 10 different tuned parameters, to solve my specific use case?

Or would I have tried 1-2 models and spent my time trying to optimize those models?

I think there is some merit to the spray and pray approach when one is in the exploration phase of the solution space.

Also, on more than one occasion now, I have had fable halve the inference latency of a model simply because the original implementation from an academic included unnecessary GPU-to-CPU-to-GPU transfers or similarly inefficient operations. Those optimizations came at essentially 0 time cost to me and I can verify that the outputs are byte-identical. Pretty sweet!

A property developer is the other real-world example which comes to my mind.

I'd be quite curious to see this combined with [transcranial] photobiomodulation - a two-pronged approach to healthy mitochondria might pay healthy dividends.

For those hearing about PBM and feeling skeptical, I was as well, especially with all the claims of therapeutic "red light therapy" circling the internet. But do a quick search in pubmed and you will see that at this point there have been thousands of studies exploring the use of PBM (808-1064nm)

I quickly found these two which relate to sleep quality:

https://pbmfoundation.org/brain-photobiomodulation-improves-...

https://pubmed.ncbi.nlm.nih.gov/41125953/

If anyone is interested in PBM I gave an overview presentation a few months ago to a lab studying noninvasive neurostimulation, I would be happy to send it your way.

I think if this were true, then individualized mastery learning wouldn't prove to be so effective

https://en.wikipedia.org/wiki/Mastery_learning

Yes this is true, but we actually have a lot more data to back this on than exclusively fMRI analysis - for example the ADHD medication guanfacine works only because the alpha 2 receptors happen to be wired differently in the prefrontal cortex than it is in other areas of the brain (a2 is inhibitory for most brain regions, but in the PFC they're positioned to amplify connections between neurons) , so by stimulating alpha 2 we allow for a more “top down” control from the prefrontal cortex than we do without, which improves executive function.

So that is one extremely robust way to understand neurological conditions like ADHD or Parkinson’s

Hi, do you have a way to get in touch so I can ask you a few questions about your lab? Thanks

If this article resonates with you in even the smallest way, I urge you to read Neil Postman's "Amusing Ourselves to Death: Public Discourse in the Age of Show Business".

I am currently re-reading this book and am amazed by the apparent accuracy of his analysis, which is that the mediums in which we communicate or express information (print vs. TV vs. TikTok) have a massively understated role in the quality and type of communication we participate in. That is, as print lends itself naturally to logical argument and less to emotional knee-jerk reactions, the type of conversations taken place in long-form print will by nature be more logical and intellectual. Compare this to TV or short term videos, which captivate us using more primal forms of distraction (bright lights with moving images, fast talking, "Gotcha" type rhetoric, cool dances, sexual/romantic behaviour, or background subway surfers), and it is obvious that the nature of what we see is inherently less based around logic and reason.

And as a consequence, if we are what we consume, it is only natural to surmise that the quality of the mind follows the quality (and qualia) of our media.

I think the prior probability in the bayesian sense is that the two entering cohorts are equally skilled (assuming students were randomly split into two sections as opposed to different sections being composed of different student bodies). If this were the case, the implication is that performance differences in standardized tests between cohorts are due to the professor (maybe one of the profs didn't cover the right material), so then normalization could be justified.

However if that prior is untrue for any reason whatsoever, the normalization would penalize higher performing cohorts (if it were a math course, maybe an engineering student dominated section vs an arts dominated cohort).

So I guess.. it depends

I'm not familiar with GPU architecture, is there not a shared L2/L3 data cache from which this data would be shared?

These look pretty fun, have you played with them much? What kind of range can you get?

yes, I definitely agree here. We've known for a long while that 1:1 therapy isn't the only way to treat depression, even if we aim to use psychotherapy methods like CBT/DBT.

David Burns released his CBT guide "Feeling Good" in 1980, which he labels as a new genre of "Bibliotherapy". His book is shown to have clinically significant effects on depression remission. Why can an LLM not provide a more focused and interactive version of this very book?

Now, I agree with you and the article's argument that one cannot simply throw a gpt-4o at a patient and expect results. The LLM must be trained to both be empathetic and push back against the user when necessary, forcing the user to build nuance in their mental narrative and face their cognitive distortions.

This is also useful for power-constrained systems - 128T x 35W is less power than 2 x 64T x 35W which gives a larger power budget for whatever h100-type device you have connected to these drives

Hey sorry to hijack a comment, I saw your comment elsewhere about the Georgia Tech's MS Computer Science being a waste of time and I was curious to learn a bit more about your experience.

I'm currently considering the online masters but still was curious to your experience if you don't mind. Reply here or an email would be appreciated

Performance gains wouldn’t be that large as enterprise SSDs already have internal capacitors to flush pending writes to NAND.

During typical usage the flash controller is constantly journaling LBA to physical addresses in the background, so that the entire logical to physical table isn’t lost when the drive loses power. With a larger capacitor you could potentially remove this background process and instead flush the entire logical to physical table when the drive registers power loss. But as this area makes up ~2% of the total NAND, that’s at absolute best a 2% performance benefit we are potentially missing out on.

> EA is diffuse and deliberately amorphous; anybody who wants to can call themselves an EA... But with no official leadership structure, no roster of who is and isn’t in the movement, and no formal process for dealing with complaints, Wise argues, it’s hard to gauge how common such issues are within EA compared to broader society.

This passage reminded me of this article: https://www.jofreeman.com/joreen/tyranny.htm

Moral of the story: be weary of groups with low accountability and vague power structures. In a vacuum, power structures will always emerge, so it's generally better for them to exist in the light than in the dark.

True, I suppose there are the "extreme" users of amphetamines which I haven't considered. Though I wonder if those who take megadoses of adderall are the same people who already are willing to consume methamphetamine or other street stims.

This isn't the same person as the student or office worker who prefers taking low dose methylphenidate to their usual two cups of coffee. If the drug was obtained illegally, many consider this to be misuse or abuse of the drug, and this is the caffeine/ADHD med cultural discrepancy I was pointing out.

So I suppose it is fair to ask: with the availably of black market stimulants, to what extent does restricting ADHD medication availability cut down on stimulant abuse?

Something I'm curious about: why do you think that an excess of stimulants is societally damaging? For the past few years I'm of the opinion that these substances should be available to those who find them useful (maybe with a doctor check-in required for blood pressure and dosage regulation).

While the abuse potential for adderall has a stronger case, something like methylphenidate has a relatively low abuse potential and has been shown to be neuroprotective (unlike adderall which may be damaging long term). Requiring people to go through a semi-permeable ADHD screening and visit a doctor every 3 months for refills is unnecessary load on an already overburdened health care system (I'm in Canada).

Why are caffeine pills ubiquitous and accepted as a means to improve productivity, yet other stimulants are not? In my mind, this parallels the argument as to how alcohol is given a pass compared to other recreational drugs since it is already culturally accepted.

Very true. I actually quite enjoyed taking the bus in small town BC, I never felt uncomfortable by the other passengers.

The same can't be said about big city transit. Just look at the demographics of where homeless people flock to.

I love this idea in theory, but recognize that you are moving the discussion from providing effective, safe, and equitable transportation towards solving homelessness.

Both issues are important, but in the context of this article one must ask, "what is best for the majority of public transit users?". It's certainly possible that a $2 rate improves quality or security meaningful ways.

The problem of "sketchy homeless and mentally ill people" is large enough that I personally prefer cycling over public transport when possible.

I think the parent comment is referring the homeless population which is violent or too public in their drug use/paraphernalia.

Pick the right bus line in Vancouver. You will notice mentally ill, violent people, and it certainly discourages the public from using public transit. Now, will waiving a $2 fee change this meaningfully? Hard to day. But it's not helpful to ignore the problem.

I'd love the link if you have it

The state has an interest in keeping its cities economically healthy, perhaps more than the city government itself which is focused on appeasing citizens with its delivery of police/recreation services. The state government can act more dynamically since it has a wider voter base to consider, including those who would like to live in the city but are pushed out.

The Valley must maintain some level of competitiveness with other city hubs, else the techxodus worsens and California loses a large portion of their economy and tax revenue. This is an existential threat to California, and the residents of the valley will lose out in the long run.