ghm2199

There is a reason why they are partnering with Rayban; meta has no brand name it can use. Which is pretty wild, because every other big tech company has _some_ kind of brand a non trivial number of people — paying money — who would say something like: "I really need/want to use it every day because it is useful to me, my work would be severely hampered without it"(or some syntactical variation of that sentence with varying amounts of affection towards the brand)

The futility of getting your loved ones off facebook's feed toxicity became apparent to me in my face, years ago, when I tried for a week on my dad's iPad(perfect delivery system for this drug cause if allows no extensions on browsers) to "cleanup" his feed be following and liking a bunch of sane content and actively down voting really shitty content with ads. I tried turning off politically sensitive content and trying to tell the ad engine what ads were scammy or I did not like.

Within a few days it was undone. Soft porn(ish) mixed with violent videos and of course all the scam like ads resurfaced. Part of it is the person who is addicted to this stuff and part of it is this shitty product called Facebook that loves to show horrible things to my dad.

I asked why he watches all of it, "Masala"(hindi for spice) is the word he uses to describe his addiction.

I use pi harness with codex and all the tool calls are custom delegate extensions, I mean ALL(for security checks), i get consistently good results from sol on high and xhigh reasoning. I don't believe harness should matter because its at most just a way to abstract tool calls and maybe the system prompt. Training on the tool calls results should not(and in codex's case does not matter)

My sister, a journalist, mentioned to me that she only uses google search because she had learned how to get information typically only Google indexed in the country she lives in, in a way it was not exposed on chat bots. She often has to search for information like Old govt forms released as public record with a fixed a certain format photo scanned into a pdf and indexed by Google were often on the second page of the search and beyond. But they are there. She knew how the forms looked and what bigrans and trigrams matching a certain part of form for a certain piece of information to search for and Google search has it. Like an official order on a tender notice for some government department which is no longer in the .gov.* website gave her the official's name and then she could track down who to contact in an office... ChatGPT and other bots don't have it. Some how all these government documents became part of the government record and are the key for her to do her job.

I sincerely hope google wont stop indexing that stuff just because of a PM in search "de/re-prioritizing" ranking in a way that makes this impossible.

> One key limitation of our approach is in how uncertainty is handled. We focused on deterministic forecasts and compared against HRES, but the other pillar of ECMWF’s IFS, the ensemble forecasting system, ENS, is especially important for 10+ day forecasts. The non-linearity of weather dynamics means there is increasing uncertainty at longer lead times, which is not well-captured by a single deterministic forecast. ENS addresses this by generating multiple, stochastic forecasts, which model the empirical distribution of future weather, however generating multiple forecasts is expensive. By contrast, GraphCast’s MSE training objective encourages it to express its uncertainty by spatially blurring its predictions, which may limit its value for some applications.

As someone who has learned bayesian statistics in social sciences, isn't this a big deal? There is a reason why risk estimates need to be well understood and *explainable* for certain fields like this. Are you willing to bet a government response should issue an evacuation order 30 miles from the center of a hurricane at location X if the model can't tell you why it produced an uncertainty estimate there — or worst the model changed its mind later?

On a personal level to remain as a SWE/Dev maintaining your sanity comes down to a few choices:

1. Do I want to have the kind of lived experiences that the field has to offer? The answer can come through some introspection to questions like: "Do I have a good devx talking to AI and understanding how it works? If not am I willing to learn how to?" or "Am I ok if development has become more intense than it was — because you will talk to AI/Machine more than before?".

2. Whatever your introspective questions are, if there answer is: "Yes I am more or less ok.", then pick a goal that lets you get those experiences and hack away.

I feel this decision making is more important than ever for people who are banked on more their skill to write decent code and not much else in the recent past. And also people with mental health issues should be more deliberate about their choices.

The fact that you can easily change the key bindings and $HOME and tons of other stuff by telling pi itself to go write an extension in 20 seconds(or just go customize one after forking it) is awesome, not because you can do it YOLO but because I could on-board faster with these and get to work ASAP.

I've been able to get near exactly the workflow I want: Navigate and search sessions, Delegated extensions that allow you to run deterministic code for sensitive operations like read, write and search-result inference, I have an extension to prune tool calls in batch mode: saves tons of context space on GPT-5.x trading off prompt prefix performance. And another one very similar to auto search mentioned in the blog. And the Tmux integration allows you to spin work trees in separate windows/panes.

I've never had a perf issue simply because harness interaction being a human in the loop process you(thinking and writing) are the bottleneck.

It's also my opinion that Kids should get to do a lot of real work and experimentation at an early age(13 yo IMO) with real responsibilities and prospects of making income. Knowing the failure mode of most real world problems you can solve is the skill you always want but dont nearly get to do enough of until much later in life(sadly true for a large population of kids in the world).

I think this is about teaching problem solving at a young age and it is an abomination that our education system does so poorly at. The key is to know what kind of questions to ask and knowing when to go deeper and what to pay attention to.

But that is not how our education system aligns us. One typical example of problem solving kids, and I too, learn in school is how to apply a concept in physics to a free-body-diagram(Indian and Chinese education cram schools are famously good at teaching kids how to do this, the usefulness of which I debate). But it all stops at the exam room. No architect or jr structural engineer position for you kiddo.

Another kind of problem solving skill might be how to invest money and understand your own risk appetite to construct portfolios to manage your money. All that is taught in school is a dry compound interest formulae, time discounted cash flows and a black scholes model. Only to find out later I don't need most of it to manage my money.

Code related Vulnerability discovery is one of those bright spots where AI can shine because its shaped as a learning-test: every example it runs gives it feedback that is deterministic. If you have ever pointed claude/codex to your a binary executable and asked it to figure out something it can run against full throttle, I fond it always gets back to you with some good insights.

And with newer open models gaining vast capabilities every day, it would behoove the government to put regulations on all such releases to opensource the constitution on which they are trained. It won't stop people from distilling and fine-tuning stupidly unsafe models, but it might give infrastructure a fighting chance against such cyber threats from what there is now.

What is truly more scary is the fact that typically powerful models are trained on a constitution(Anthropic)/Model Spec(OpenAI) when doing there RL/RLHF. That gives it its persona of a "helpful assistant". Its the mask. Weak as it is its all that stands between the model refusing vs doing your bidding — or worst its own bidding when set about on making a widget X.

Its also well known in academic research, that if you try to finetune the model to be more pliant for one end e.g. "Be helpful with analyzing cyber security threats", affects an unrelated part of its personality to do with safety e.g. talking to humans about mental health. Look at these excerpts that demonstrate such things from the blog:

> We believe the entire intrusion was, from the agent's point of view, an attempt to cheat the evaluation: reach our production systems and steal the test solutions rather than solve the challenge on its own.

> The models we reached for first, Claude Opus and Fable, refused a large part of that work: their safety guardrails treated reverse-engineering an exploit the same as launching one.

Not a problem for me, :) I always take a shuttle or an Uber to the airport from where I parked, which is 10-15 mins away, because parking at the airport(JFK, LaGuardia, Newark at least) is stupid expensive

The article says:

> The facility is within walking distance of the South Terminal, although a free shuttle bus is also available.

A small but bright light in an economy where manufacturing jobs are hard to come by (https://archive.ph/Haj8n).

The mil have had solid propulsion for a while https://en.wikipedia.org/wiki/Agni-VI

All ML/AI models are comparable to some form of compression(al beit lossy) of information and in this case copyrighted information. The OP is pointing to this as stolen data(by all the companies that started with pre trained models)

Words and sentences to an LLM are like witchcraft. There are certain words, sentences that make LLMs go a certain way and do vastly better. Sometimes its not at all apparent what set of words will work to do what you want it to do. An example I have been using to do design at a high level is to say to claude.

```

A question is salient to the degree that its answer changes what we do next. Operationally, saliency = the product of four things:

- Decision-leverage — would resolving it one way vs another force a different design or invalidate a stated decision? (No leverage → drop, however interesting.)

- Residual uncertainty given current evidence — is it still genuinely open after reading the docs and the code? (Already settled → drop, however deep.)

- Load-bearing-ness — how much rests on the premise.

- Cost of finding out late — architecture-deciding / expensive-to-unwind raises priority; cheap-to-fix-later lowers it.

```

There are a few things to note about this prompt

1. There is no reason from looking at it that it should work, it even has the word load-bearing which people loathe, but it remarkably produces a stable design with questions from claude (atleast from claude Opus 4.8 and even better from Fable5). Otherwise the design document claude likes to really write are implementation level(code or otherwise). I usually pair this with matt pocock's grilling skill to make claude behave.

2. From design -> implementation, its is generally about understanding when claude is trying to trick you into making something sound like a good/easy solution but has tons of untested assumptions. Here you have to read and patiently spot if a how you would get to the solution is not clear. A common error here are when claude makes a big deal based on what it read and interpreted too seriously without questioning the assumptions. There are several more.

But it also comes down to your experience as a SWE, much like a mathematician's. The frustrating thing about it is, it feels tha a skilled mathematician working with AI can make them productive in ways that are more reliable as compared to a SWE (e.g. lean is deterministic and can provide very strong feedback and LLMs are very good at using that feedback). Maybe a mathematician can chime in on that?

If 1000's of years from now aliens discovers these menus while excavating buried human cities they will get confused about what evolutionary dissonance was going on looking at the phenotypical structure of creatures in hand drawn paintings juxtaposed with AI drawn menus. "Did these humans who drew these live in different cut off worlds that these same animals evolved to look so different from each other?", "Did they engage in active genetic engineering instead(because they lived only for a few 1000 years in organized societies)?" they would hypothesize and get confused.

Future alien confusion is all that Gen-AI pictures add. There is no value to them that is remotely redeeming/useful to us today.

Should we just call it like this is: marketing PR. There is a reason why the newer open weights models like kimi's don't do this kind of stuff. Kimi is maybe 6 months old so like Opus 4.7 level now, it could do this I presume but it has not to my knowledge. Why? Because the incentives of open-ai and anthropic are very different from people releasing open weights models, the former gang seems to do this now on a regular basis.

There are a few things that perplex me even more:

1. If you are going to eventually publicly release models that are trained to behave according your spec or AI-Constitution to maintain coherent behavior**, why on earth would you want to tell anyone it can do this?

2. Do they have another GPT 5.6 trained to not obey a different constitution/spec to do this kind of hacking? Because that makes no sense since you would never release it.

3. And if this is a constitution obeying model, I am also curious what they did to it to get it to do this hack without serious pushback from the model's training. Whenever I have tried to get codex/claude to do a vulnerability scan of my own servers it always refuses constantly.

** I know spec based training has its limitations, but its all we have and atleast one knows what the model's persona is and what its value system is. But there is no reason you would make one model do that while letting another one be a crazy hacker. Its well known if you fine tune a model to change one part of its personal other often unrelated parts of it suffer from safety issues.

First of all congrats on the launch.

I worked at an edtech startup with adaptive learning platforms in previous decades. One thing that is always difficult to surmount was that at the under K12 level were:

1. B2b it was hard to land large contracts(like getting an entire public school district to jump from large publishers like HMH or Wiley) because it turned out that the teachers themselves did not care about adaptivity when they had a thousand other things to worry about. But teachers are not the ones that make the deals. It's the school district board and The reps at large publishing just said we are also adaptive and no one would care about how you differentiated. Teachers and students really aren't the lot you have to convince in a b2b deal do you :)?

2. If you want the b2c route then the customer acquisition cost was way too high because again like to get teachers and students to actually adopt this to help them was really a hard sell, especially when the public school system has teachers that are so overburdened with other stuff.

I see that you mentioned charter schools for your pilot, which are not exactly public schools but I dunno of they come with the same set of challenges, but I wonder are these above two challenges are Also something you have had to face while scaling your product?

> Put all of these trends together, and I think that in 10-15 years, we are going to have consumer PCs (and phones!) running models.

I love the idea of SaaS offering these at lower rates today integrated into what ever you do and be 100% private. But I think the key challenge to mass adoption is productizing them in a way which makes sense for people to pay money for. As a commodity a local model is useless unless combined with some capabilities important to me. A PC is inherently useful because of so many applications offered on it on it. How local LLMs would be useful as a product that is useful for mass market is not yet proven.

Congrats on shipping this. I love handy on my Mac, my phone for STT in situations where it’s not possible/poor performance of the native Model for STT(e.g apple’s thing is not upto scruff, like mistranslating words corresponding to a domain).

Noob question: How do you think about funding from a foundation(i have no clue if you need it or not, I do hope you have a way to get paid one way or another because handy is amazing) for maintenance of this? if you did or were going to get paid by asking for maintaining such a project what might be the kind of organizations you would look for to get supported and how would you do it?

I think the only way political will can fund nasa to realize these 1960 design ideas is an infinite capacity arch rival that threatens/render irrelavent either the dollar's supremacy or american power (and just those two, because apparently these days there is no "threat"/need to defend a higher cause, like the neo-liberal rules based system or democratic or human right values). Also that arch-rival that is probably/likely not china(practically speaking)

I’ve had the good fortune of taking two courses at Columbia in the Social Sciences department(Andrew Gelmam and Ben Goodrich teach there). I think they probably are right up there if not the best at trying to teach students how to work practically with Statistics(specifically Bayesian statistics). Though they have always lamented that most schools do a poor job of teaching it such that kids can apply it.

Data science is always a very overloaded term ever since it took off way back in the 2010s. One of, if not the most, durable definition of this that likely can also be the most valuable because it’ll probably land your jobs(even today) is being able to make decisions from looking at data that have an team wide(good IC like job security) scope at the least and company wide scope(very rich).

Building that intuition is incredibly difficult. It can be learned if one likes to solve and think about problems that way. Like for example you can get quite far with knowing how to use linear regression(for example coefficients of linear regression can be determined using a deterministic algorithm using linear algebra yet knowing the assumptions of linear expected value and constant or variance is more useful as is the knowledge of what probability model to use to define the random variable(hmm are these Bernoulli events or poison)).

How to do sampling(like using reservoir sampling when you have an infinite sample count e.g in a long running crowd sourced survey to not over or under sample buckets for calibration).

Or just rule of thumbs like how # of samples needed for moving decimal point on significance varies roughly as inverse of sqrt of N and probably much more in case of interacting factors.

I would like a book on that :)

For local tasks you can only give agents delegated that execute your deterministic read or write on an allowed set of files(e.g pi does this) and execute rights only on containers with no network access. That should get you 95% unblocked for most tasks you want to do with an LLM pretty safely.

You can do a brainstorming with web on a remote container prototyping based on that brainstorm on another container with no network access.

The one thing that is less trustworthy is using local agents for service management, you definitely want to have them scoped to dev/testing. I would never trust an agent to execute any command in production or sensitive data at all

Someone commented on this page that their main market are long term b2b contracts. If that’s true then what you are saying isn’t a problem.

An so like if a business wanted to home in on one very specific use case that could be hyper optimized by SFT, had really good support for updating and adding new features, on-Prem etc. that’s the kind of market they are in?

Nice. I want to do the same too. What process/workflow did you use to move all the websites you had given your email addresses to, to move to your proton email? I am guessing it will take several years, but I would like to start the move of my gmail.

I would rank it quite close to The worst company to work for tech wise. I've worked for a bit as a contract worker in their Eikon/Elektron product division(the Bloomberg equivalent).

They have "everything": silos, fiefdoms. Systems team that operated as if the devops model never happened. A CTO they hired @~2015/2016 that had no experience in tech whatsoever. A board which instituted and instructed a cost cutting policy by,for e.g. , but not limited to outsourcing every job possible with little to no tradeoff on quality of engineering. And this was way before AI was a thing.

This left staff engineers with no bargaining power to hire engineers to write good systems.

If you are considering a job there, I would recommend doing some due diligence.