bob1029

Knowing what I know about publicly traded companies and their willingness to bend roadmaps into contorted pretzels, this is not how an "imminent" RSI capability would be advertised to the market.

If it was truly imminent they would claim they already had RSI. That is how the game works. Take a look at IBM and quantum if you require an additional sample. Review the things these companies strongly imply they can do relative to what they can actually do. Anthropic is arguably making less aggressive claims with this RSI narrative.

The lies flourish in the definitions behind these terms. When Anthropic says "RSI" and IBM says "Quantum", they are invoking a highly curated lawyer routine that has a mountain of legal documentation surrounding it.

If we are being slightly reductive, this is just more of the exact same thing. The token stream in this case happens to be a projection of the weights of a neural network. I don't see where we violated any of the prevailing assumptions about how these systems operate.

The mapping between spaces in this case is mostly linear. This isn't like gzip or a Turing machine. I'd argue this is even more permissive (easier) than the LLM's originally intended use case of human language.

What happens if we can come up with a model that can directly emit a frontier LLM given some seed context? I feel like we could train a model to recreate astra verbatim. At some level it's just bytes on a disk that can be bundled up into tokens. As long as the representation is rational and consistent I don't see why this wouldn't work.

Why not use a completely different VCS topology if you are a gigantic trillion dollar behemoth? Why are we still using the "distributed" git concept as the basis for everything?

Centralization is exactly what you want at this kind of scale. The most marginal participation from the server would make a huge difference.

There is a reason game studios don't use git. It's not just about the large objects.

> I'm not sure how AI could destroy the demand here.

By overwhelming the supply. How can a consumer keep track of what to play anymore?

As an example, I didn't even hear about PBZ until you mentioned it just now. The fact that Valve might occasionally select lottery winners is separate from other forces involved. I would hesitate to highlight the occasional breakout (artificially-promoted) success as some kind of evidence that there is nothing to see here.

> They just need to make good games.

If only we could just do that.

I think the best answer is AAAA+ scale projects. AI has destroyed the demand for something that can be produced in a few months. The "meso" scale of game development is extremely difficult to work in. Constantly walking a tightrope between too easy and too hard. Rockstar games does seem to have the right idea. They can remain viable by simply never reducing their feature scope and spending unlimited amounts of money.

Perhaps enough indie game studios could align their thinking at the same time to accomplish a similar outcome. In any case, pooling of talent and resources is probably the only way to compete now. The individual developer simply cannot keep up despite being able to do more than they ever were before.

If you seek evidence of Jevon's paradox I think you may find it here. Albeit in a very undesirable form. Even if I can now do 100% more work per unit time, my prospective customer is demanding 300-400% on top of that in terms of the final delivered experience. The water level has gone up everywhere and continues to rise rapidly. Even outside of game development. Deliverables that would have elicited reverence from the business twelve months ago are completely unrecognized now.

It's not hard to find options if you stop looking in stupid places with bad search terms. "TV" on Amazon is literally the innermost circle of consumer hell.

> I don’t mind paying a premium

https://www.bhphotovideo.com/c/product/1916990-REG/lg_65ep5q...

> Since ChatGPT chats cannot spawn or control subchats/subganets so I had to be a bit creative in order to do multi-agent development and orchestration.

Multi-agent is mostly dead. One big agent loop running GPT6.1 absolutely dominates any weird ideas like this now. I still use sub agents for some tasks where I want to compress the action space and context. For example, testing a workflow that requires 100+ tool interactions for each logical test. Without a subagent it's really hard to guarantee a test makes it E2E. I have some very "persuasive" methods for long horizon tests that would break the main loop's dynamics.

We may not feel so compelled to exploit the system if we could figure out how to not be so wasteful. Multi agent schemes can bloat token usage by orders of magnitude. They force us to consider consuming less ideal tokens from contexts and vendors with higher variance.

HN loses its marbles when I suggest it's economically feasible to prepay for API tokens and use these as the exclusive means of performing AI work. I get done with a thousand tokens what my peers cannot do with a million. It doesn't matter how many hammers or workers you have if you don't have the aesthetic preferences and domain expertise to employ them correctly.

The biggest issue is any wind damage to the roof. If it's leaking even a little bit, it's going to be leaking hard for hours.

With a consumer dehumidifier you might get 5-6 gallons per day. A 5-ton HVAC system will pull over 50 no problem. You might have hundreds of gallons of water soaked throughout the framing, drywall, insulation, etc. It can take a really long time to remediate without a powerful HVAC system.

> For the vast majority of the population

In the spirit of nuance, the vast majority of the population of Florida is at risk. Something like 10% of the US population in total could be impacted directly. You don't actually have to live on the coast to get rocked by a hurricane. Ike took out power for millions in Ohio.

https://www.facebook.com/100044470387919/posts/ohio-remember...

I found this to be unusually good. I got caught in a rabbit hole about interpretation of the work of Bach. I would have never seen this stuff in the YT interface.

My biggest problem with batteries for grid outages is that the only outages I am actually concerned with would drain them to zero before the outage is resolved. For short term, smaller/portable batteries are fantastic at bridging a few hours. Plug the wifi & fridge in and you can sometimes go for 2-3 days.

I am not aware of any economically feasible way to run my 5+ ton HVAC system full-time on batteries after a hurricane just reset all the infrastructure, and you do need to run it full time because of the insane humidity and latent heat that follows a tropical system. Doing the math, I'd need something approaching 800kWh of storage to ride through this kind of event (7+ days). This is approaching half a million dollars in batteries. I can install an extremely nice 1800 RPM generator for ~15% of this cost, which would also include a year's worth of propane buried on the property.

I think a simple generator on an ATS remains undefeated from an engineering perspective. If you have an air cooled unit, the maintenance is not much more complicated than with a lawnmower. You just have one extra spark plug to deal with, and sometimes the oil pressure sensor needs replacing. I can get my stubborn Honda generator to start with all kinds of weird tricks. If my Ecoflow battery starts to act weird, I have very little recourse by comparison.

I'm not sure I agree with the conclusion that this is merely caused by UPS system control logic. That is an extreme amount of coordination across a lot of makes, models and configurations.

The FFT indicates this is a periodic signal, not a once a day or week thing. It's happening continuously all day every day. I don't know of any UPS systems that perform tests with this kind of regularity.

In any case, I refuse to believe an external sync signal is causing this. The more likely explanation is that the active elements are pulling each other to a common mode due to being physically coupled on the same grid. This is like how clocks on the same beam will synchronize.

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

The guardrails on Anthropic have become an albatross. I've got OAI models doing testing of banking workflows that Claude would have a full scale panic attack over. Even OAI needs some alignment and reassurance regarding the fact that it's not about to pull credit for a real person. GPT6.1 will happily launch the nukes 100% of the time if you make a clear and rational case in your system prompt.

Forcing this kind of safety into the models and harnesses is a mistake. I think the strategy of post hoc moderation is the right one. Trying to prevent evil ideological activities as they transpire is a cursed mission. Let someone who is genuinely doing a naughty thing accumulate a meaningful body of evidence and then send in Mr Holmes (the FBI / police). Attempting to prevent wrong think at every token is too granular. We have systems designed to handle the possibility that crime and punishment may occur at separate moments in time. Most people don't go around murdering because the eventual consequences are pretty severe. Just because an AI gave you advice doesn't magically make those existing consequences vanish.

> but it also has strengths too (e.g. computer use / control; which is great for automated E2E QA).

Using 6.1 Terra for E2E workflow testing feels like cheating. I thought it would have all kinds of problems with the action space on certain steps, but it runs like a deterministic unit test.

I am convinced that OAI has the superior models for most practical scenarios. I've tried Claude recently and it makes me wonder if some of yall aren't actually massive trolls. It's genuinely not in the same room of performance if you place one foot outside junior programming or shiny blog posts about rewriting factorio to run on a ps2.

The supply-side is fine. Even on Spotify. There is more content than I can listen to. I go out of my way to favorite things that are "marginal" so that my recommendation profile stays diverse.

The demand is what is collapsing. 30 seconds of trashy TikTok music here and there is about all the average consumer demands these days.

At some level this all just comes off as post hoc justification for a really bad cortisol addiction.

Why are we keeping score in a game with completely arbitrary rules? The polarity is the hallmark sign for me of someone who enjoys a ride that certain neurotransmitter. I used to commute to and from work 2 hours every day. I know what it's like to be addicted to this stuff.

I think a lot of companies are struggling to move beyond this being a thing developers use to something everyone can use. Developers have to go really far out of their way to build agents and tools that apply the capabilities to their specific business. Someone in accounting cannot simply fire up chatgpt.com or copilot and use it in the same way a developer can. There is no instruction manual or historical precedent for how to do this. Very few people are currently in the right situation to make this work.

My observation is that most development teams are still in panic / survival mode, spending all day stacking sandbags instead of bringing these capabilities to the broader team. The fear of job loss in this particular market is crippling. No one will admit it publicly but you can feel the adversity on every AI status call. I've managed to convince a few that AI won't take their jobs, and these people have become indispensable at enabling deployment of AI beyond the weird nerd part of the org chart.

We can break the 5500K temperature limit with passive elements and some clever engineering.

https://en.wikipedia.org/wiki/Solar-pumped_laser

https://ieeexplore.ieee.org/document/1553831

Coordination is a tax. You don't want to chase this.

Everyone who is doing community of models via message passing is losing the game right now. Every single case of gas town I've seen could be refactored into a single agent loop with a well curated set of tools.

I think designing the tools is the hardest part in doing this AI harness/agent thing correctly. If you feel inclined to create an entirely new role in the system, it's likely you couldn't figure out a clean way to make two tools work together. This conflict is not resolved simply by moving the two tools to two different agents. This merely hides the old problem inside a new one.

This is the first time I'm reading a rational take on distillation.

If China does not develop their own frontier soon, they might never be able to. If US labs get any faster at releasing models, china will not be able to extract the value from the current generation before the next generation arrives. Distillation does not happen instantly.

HN might be content with "good enough" Chinese models, but my clients certainly are not. They ask the robot questions that challenge the very limits of comprehension and now have some pretty high expectations moving forward. HN asking mr deep seek to refactor their shit pile is not a high bar anymore.

> What's the threshold/requirement for that to happen?

The RoCoF (rate of change of frequency) must be low enough that the inverters can form the grid without letting out the magical smoke. Having inertia is what slows this rate down.

This sounds an awful lot like how semiconductor manufacturing works. I see direct mappings to FOUPs, AMHS robots and reticle pods.

FOUP = thing that holds the PCB.

AMHS = thing that moves the PCB.

Reticle Pod = The thing that holds the component(s) that can add value to the PCB.

You can create a much more capable factory if you trade a fixed, serial line for a scheduling and logistics problem. Assuming your logistics and robots are capable enough, this can provide enormous uplift in high-mix, low-volume manufacturing.

For example, running a test lot (up to 25 wafers) through a modern semiconductor plant is almost a trivial affair. It is like a leaf floating down a river if the scheduling systems are well engineered. I suspect CircuitHub is attempting to leverage similar dynamics.

I'm still waiting for one of these companies to provide the actual hard part.

The web front end (chat UI) that the business actually interacts with often needs to be adapted to the domain. We have tried to use the various chat UIs out of the box with customization on top and all of this has failed spectacularly. There are cases where we need to pop into an inline approval flow with custom elements and controller actions. How do you hook the existing platforms for this? There zero path. MCP can give you an information channel, but it does not enforce the required UI/UX flow for certain tool calls. We could either compromise on UX or rebuild chatgpt.com from scratch. We ultimately went with the latter.

If only we could discuss the OLED display rather than an optional feature. I did not see anything in the marketing materials that indicated I would be punished for refusing to utilize the touchscreen.

I think half of y'all are trying to justify your most recent MBP upgrade. Those of us still on the M1 are probably looking at this thinking about how fortunate it is that we've waited this long.

The OP wrote:

> capital punishment for such gross negligence

"gross negligence" is a very specific legal test.

> Gross negligence is a lack of care that demonstrates reckless disregard for the safety or lives of others, which is so great it appears to be a conscious violation of other people's rights to safety. Gross negligence is a heightened degree of negligence representing an extreme departure from the ordinary standard of care. Falling between intent to do wrongful harm and ordinary negligence, gross negligence is defined as willful, wanton, and reckless conduct affecting the life or property or another.

https://www.law.cornell.edu/wex/gross_negligence

It's expensive to move humans around. Sending bits across the internet is very cheap.