itissid
In a muddled political system(democratic or otherwise) today its about power and how is it balanced. When Nandan Nilekani was creating India's UIDAI(the national aadhar system), he insisted that he would only do the project if he was given a cabinet level post[1]. His is one of the few technology examples to succeed in India when most large government run projects are a disaster. ISRO is probably another decent example of success. I think Nilekani succeeded because he had a clear moral compass and a defined goal.
But there is a problem with kings, soon they have an asshole child. Metaphorically speaking at least. UIDAI is being used stupidly enough like an SSN now and in ways it was never meant to be.
[1] https://www.newyorker.com/magazine/2011/10/03/the-i-d-man
If there was an inflection point to break that perception — like covid gave a pause for thought for location based ad targeting — and hence the incentive, perhaps we can kick email as one of the ideas whose time has come.
I am rooting for ATproto to be the replacement.
[1] https://www.glassdoor.com/Reviews/PROTON-Reviews-E13574.htm
1. Ask them about what percentage of their own resources goes towards cybersecurity, audits etc. Would they even tell you.
2. Technical stack maturity. If the details are available, because they won't reveal any of it to you.
3. Other signals: Do they give talks on platform security, who is incharge of their cyber security.
if I understand taking LLMs as an example certain neurons will fire based on certain motifs, then one can change the phrases or sentences around that motif and _predict_ if the same neurons would fire.
Like if I said "I like cars, but I like rockets better. Can you suggest some rocket models to fly?" and Neuron X fired every time it saw a `, but` and attended to what followed it, that would give you a good way to _understand_ when it changed and what might have caused it.
My understanding is that this is now clearly visible behavior.
If its grouping circuits based on motifs and when they fire — and it applies to large networks — then one should be able to literally query or debug a reconstructed database and check against actual activations. More importantly If this was a hand-wavy thing to do earlier for safety, it's not anymore because it's on more solid ground using Normal Science.
Many people don't even know what locking in means. The crowd in this forum < 1% of the world do, you and I do, and maybe a few outside. There are only laymen outside of 99%. No one i know of knows or uses a harness. They don't know the large number of VMs built on say Hermes. They really don't care or have time. If you give them a ready made product thats a click away, and its good enough, they will use it.
I am scared for my parents to get exposed to this kind of thing from bad ad network actors. Does anyone have some practical concrete recommendations they took? Like I don't think they can install ublock origin and whitelist sites on demand in Firefox they way I do.
The only practical thing right now is to make sure they don't have the money on device and have 2 fac for email and bank accounts. But every now and then there are phishing attamepts that they could fall prey to.
I think what they are missing is the fact that there are not that many teams in the world that can deploy muse like product to A/B test a lot of hypothesis at the scale meta can at scale. I mean even if you could get to scale with DC you need a reason for people to use your product.
My mom and dad and their family have never heard of Claude and don't use chat gpt except once in a while as a search engine. But meta will figure out a way to insert this into modals and feeds to make people stumble and stick on it.
I am no meta fan, but I think people underestimate how fast product uptake there is for non Dev harness.
Meta being meta, will likely run this as an "experiment mode" for 6 months to an year and if it does not work scrap it all and move onto the next thing.
There is a risk ofc muse does not work. But there is a good chance that with the ridiculous number of product teams doing A/B testing with insane number of people already on it, something will emerge that is interesting in a way it never could for their erstwhile Ar/VR foray.
1. First several 3-d design iterations are throwaway reworked because I am constantly measuring the design, clearances and studs. And paper works best here instead of working with clankers which takes way longer.
2. Choosing the right kind of materials: thickness and type of plywood, type of hinges, panels, finished or unfinished. Optimize for simplicity of putting components together is most important.
3. Once the design settles the actual joinery needs to be as simple and efficient e.g. butt joints work well most of the time. A good jig for pocket holes is generally the most needed. You will be surprised how well a staple gun works for creating back panels of cupboards.
Its not that they are not useful, shopspec looks great(good job variodot), its just that I can get further and faster without them for most hobbyist pursuits. I invoke YAGNI most of the time because I constantly ask myself one question: "How can I make this be less time consuming?".
The trade surplus of Taiwan, China, Korea, Japan etc including chips has been sky rocketing. This is like Bill of materials(BoM) and it's at least 3-4 trillion in chips for AI alone, summing it up over the past 3 years.
If companies that are basing X% of their future earnings on new AI based products don't pay up a bunch of those data centers are not gonna be re-leased. It's gonna be a problem. An seasoned analyst can take educated guesses as to what that X is: investments made in pure chips by Nvidia/AmD vs how much capital investment— not VC — Faang(i.e ones that are just using the chips for building software) are putting into them. The difference between investment devoted to pure DC build out minus what they are spending is a good proxy to understand the overinvestment.
I think the only way meta can care about privacy, if any, is if it becomes socially taboo (lawsuit, WSJ/NYTimes report etc).
The clearest quantitative evidence of this is their P/E, its lower than others relative to its forecast earnings because of all the legal risks that need to accounted for.
Compare the two yourself and judge(latter is written in gpt-6 luna).
> One client is one API call. It comes back as an always-on Hermes agent that holds that client’s data, that client’s identity, and a hard ceiling on what it can spend. It answers on an OpenAI-compatible endpoint from the moment it boots, so integrating it was a base URL, the same as pointing at any model API.
> Boon provisions an agent for a client with one API call. Each agent has its own client data and identity, along with a spending limit. It exposes an OpenAI-compatible endpoint, so Boon can connect to it by configuring a base URL.
The cool thing is even though this tool addresses a few, use case for reducing cognitive load, they happen so frequently that they add up.
I think they key with cognitive load is that the agents often produce a lot of trace/docs etc, in the end only a small % of the traces really are important to the final changes, because concise changes are typically very small and self contained.
One thing i see this becoming important with is maintaining internal tooling built using this. I maintain an internal docs system that helps me do designing before i build code, that docs system is completely vibe coded and i can add features to it very fast, but now its all grown up. A challenge then is can i understand just enough about a new proposed change to approve it? That is key to the doc system not becoming a burden in of itself, while maintaining a tight core feature set.
To add to this the more difficult problems in classification done on scale have always been about collecting "good" -ve examples, enough data to calibrate on every confidence interval and debugging outliers. And those are solved on a case by case basis by the company pursuing its own peculiar version of the problem.
Am I the only one who thinks this is just all hype?