IanCal

11 ms·

151.
by IanCal 3mo ago
If you’re aware the person wouldn’t want you to do that, yes it’s wrong. Being able to do something is not the same as it being right to do something.
152.
by IanCal 3mo ago
> it's no different from a person who doesn't understand that their neighbours can see into an open window at night. And standing out in the street staring through with binoculars is still wrong and creepy. > Should Shodan b
153.
by IanCal 4mo ago
Debit cards are also mostly visa/mastercard.
154.
by IanCal 4mo ago
I would also recommend the approach of using an llm to create the examples, and then train from there. You can even get fancy and do things like active learning with the llm taking the role of the human annotator and sending in trial statem
155.
by IanCal 4mo ago
Are there ways of doing age gating while preserving privacy? For whom? How many people need that kind of privacy, would it impact lots of (say) teens seeking help or is this about whistleblowers? Are many under 18s using VPNs for porn? Or h
156.
by IanCal 4mo ago
Can’t you? They’re taxis right?
157.
by IanCal 4mo ago
Which models were you using under this? If you used the quality default as exists in the interface, it makes sense that it was ~4x the cost as it'd be 3 frontier models judged by one of those. The idea would be to use fusion with simpl
158.
by IanCal 4mo ago
I’m a little lost though this seems like it could be fun, on safari mobile whatever I build keeps losing some of the connections as I tap on other things so it’s hard to get far with it.
159.
by IanCal 4mo ago
> software development, as a “decide-execute-deliver sandwich”. AI compresses the “execute” layer — the middle of the sandwich — but the other two layers resist automation in a way that will not be overcome by capability improvements alo
160.
by IanCal 4mo ago
> provided you're willing to label 200 or so images A quick note to say that this is also a task you can hand to things like gemini.
161.
by IanCal 4mo ago
They can however be extremely useful for curating training data. Also things like SAM and the DINO (/grounding dino) models. Also if they are better then you can also have a flow that’s cheap model -> marginal cases go to more compl
162.
by IanCal 4mo ago
This is an interesting read: https://ai-2027.com/ I'm not going to say it's a perfect prediction, but I do find the trajectory of "can write something reasonable" to "oh can write snippets of code&q
163.
by IanCal 4mo ago
> If we agree weights producing text may emerge consciousness, given large enough, then DBs must've gained it long ago. If we agree that silicon can perform calculations, then beaches must have been working out log tables long ago.
164.
by IanCal 4mo ago
You can't do that for me either. Why would you refuse to use a patch that deals with a valid PoC exploit? If a random contributor posted an explanation of an exploit, showed it worked in an executable way, presented a patch and you cou
165.
by IanCal 4mo ago
> 98% smaller in terms of active parameters (since it's a mixture of experts model). I don’t think that’s right, this flash model is 5B active params. Qwen3.6-35B-A3B is 3B so 40% smaller.
166.
by IanCal 4mo ago
51% does not mean it randomly gets things wrong half the time. These things can be useful if you can accurately predict which tasks they will reliably do, and which they will usually fail on. Then you can get much more reliable work from th
167.
by IanCal 4mo ago
Why would it not be reproducible?
168.
by IanCal 4mo ago
While I agree it seems obvious you can’t do that, based on how these platforms have limited things for a long time… but that really should be something you can do. Why can’t I get the file and put it on another device? Why can’t I burn it t
169.
by IanCal 4mo ago
Doesn’t it flip around for small scale? Paying 100x the cost for something, all in, it’s cheaper to rent for small workloads like 10m/day. At 10x you have to be at hours per day and 5x you’re at 4h.
170.
by IanCal 4mo ago
Is latency of the network that noticeable? Aren’t we talking low hundreds of ms at worst here? Much lower for something close regionally.
171.
by IanCal 4mo ago
The cost of sharing recordings of video you collect anyway is near 0, and this provides videos of public places . What's the value of a recording inside my house to the police? That requires paying a human to go around recording it?
172.
by IanCal 4mo ago
> Is that a problem? Yes! Are you totally missing some countries, did you set it to try and hit 20% but it’s actually 60% of traffic, does it happen to include some VIP type users, how does it impact the thing you’re trying to measure wh
173.
by IanCal 4mo ago
Almost, but not entirely, quite unlike the truth.
174.
by IanCal 4mo ago
This gives you a distribution unrelated to active use, puts users in the same bucket (with the same number you’re going to have the same users in the first 10%) and links combinations together. Often problems are more complex than they seem
175.
by IanCal 4mo ago
> No developer or SaaS should be needed to make them understand tasks are checked out when you start work and contain instructions and when you are done you move the ticket to DONE. However jira, trello, linear and basecamp are all Saa
176.
by IanCal 5mo ago
A few scattered thoughts but a board with decoration or art of a similar size could be a nice cover, the other (more building required) would be to look if there’s a way you can fold down/away the monitor when not in use.
177.
by IanCal 5mo ago
Are we a long way away? https://chatgpt.com/share/e/6a0bf28b-e198-8012-9a88-c777d965...
178.
by IanCal 5mo ago
Also things like improving devx - nicer log analysis, speeding up test suites, auto handing some CI failures, improving scripting, tooling etc.
179.
by IanCal 5mo ago
I wouldn’t discount the value of moving small tasks away from developers, nor the value of fast cheap prototypes. Product owners can very quickly get, for many problems, an interactive demo without coding. For lots of problems this can be s
180.
by IanCal 5mo ago
They can ask, they can do a back and forth and they can write documentation to be used from that point onwards and write it in a common style and structure. These are language models, being able to talk through something with them and have

More ›