Your cold-outreach emails will be flagged as spam by some people, even if they comply with CAN-SPAM. And you should comply with it by having a clear opt-out mechanism and your physical address in the footer, but some people will flag your emails regardless.
I've had this happen in my startup. It hurt the deliverability of our transactional emails (notifications etc.). I'd go with a separate domain if I had to do it over.
What an unpleasant attitude. People have emotions. If they're apologizing, maybe they feel bad. Accept it and get on with your day. A punctilious email etiquette isn't going to improve anything.
I don't know, it seems pretty light-hearted. If they sent this directly to someone in response to an email, then I may agree, but since it's more of just an opinion blog piece, I find this to be a good outlet for thoughts to share without really impacting anyone.
I actually really liked the post. I'm often prone to apologizing, thinking that it's a social expectation, and the post made me smile and relax a bit, thinking to myself "oh, maybe it's not that important, and it'll be ok if I don't".
I think there is a cultural gap. He mention not only apologizing but giving explanations.
For instance my observation is that people in the USA will tend to give you a lot of unrequested information like all their health/medical problems or the sports games of their kid and what not. People in europe seems to be more private unless they are talking with very close coworker they would consider as friends.
I agree that the apologies tend to make one a little uncomfortable. This is because people do not simply say, "I am sorry I did not have time to write sooner. Here is my response", but instead say, "I would have written but my child was sick, etc.," so you feel the need to respond to that, and feel bad for having bothered them.
Almost nobody writes, "I am sorry I was scrolling Twitter and Hacker News while ignoring my e-mail. Fortunately, I have stopped and now can respond!"
I write no more than either, "Apologies for the delayed response," or "Apologies for the delayed response, I've been out of the office unexpectedly." Very business-like. Tone neutral. Easy to digest and move on.
> But if you haven’t used specifically Opus 4.5/4.6 with specifically Claude Code for at least an hour, then you’re in for a real shock. Because all your complaining about AI not being useful for real-world tasks is obsolete.
These hyperbolic takes from Steve are wearing thin.
It wasn't my experience that Opus 4.5/4.6 was a sea change. It was a nice incremental improvement.
> And unfortunately, all your other tools and models are pretty terrible in comparison.
Personally, I like Copilot CLI. $10 a month for 300 requests. Copilot will keep working until it fulfills your request, no matter how many tokens it uses.
Calling all other tools "pretty terrible" without specifics reminds me of crypto FOMO from the 2010s.
> How do you know one party isn’t 15 when the other is 25?
You don't. That's why parents need to be involved in their children's lives.
CSAM is the easy excuse, anyway. That's the one lawmakers use, and most people are against CSAM, myself included, so the excuse goes down easy. But the impetus they don't talk about is monitoring and control.
The answer isn't to destroy privacy for everyone. The government and these corporations don't need to know what you're doing every second of the day.
> That's why parents need to be involved in their children's lives.
Can't, aren't, look at iPad kids, won't. This is about as logical as saying people should just drive safely, so we don't need guardrails and seat belts. Or saying parents should always watch their children, so we don't need age verification at the alcohol store. Besides, it's not like the school library or the friends of friends don't have devices themselves you as a parent can't see.
Parents should not need to be tech experts or helicopters to feel their kids are safe online. That's fundamentally unreasonable. In which case, privacy and child safety need to come to an unhappy compromise, just like any other conflicting interest.
For that matter, I'm surprised that HN automatically always accepts the "slippery slope" fallacy while lambasting it everywhere else.
> This is about as logical as saying people should just drive safely, so we don't need guardrails and seat belts.
This is a terrible analogy. Regulations related to driving only apply to drivers, if you're a pedestrian then you're not subject to basically any regulations that licensed drivers have to abide by. On the other hand, internet regulation like this punishes absolutely everyone to safeguard a small group, that being parents. It's like legally forcing pedestrians to wrap themselves in bubble wrap while outside so the careless drivers who couldn't behave don't dent their cars and get hurt when a pedestrian flies in their windshield, when they inevitably collide with one of them. Why is any of this their responsibility?
The fact that there is absolutely zero effort in pursuing any non-punitive options (like forcing ISPs to put networks of clients with kids in child-friendly mode, where the adult has to enter a password to temporarily view the unrestricted internet on their network, which should cover 90%+ of cases; or doing any of the proposed non-identifying proofs of age, like a generic "I'm an adult" card you can buy at the convenience store) should tell you that this has very little to do with actual concern for children. They went out of their way to enact the least private, most invasive, most disruptive option, which will not even work better than any privacy-friendly options, unless you expect literally every website on the internet to be compliant. Teens are smart, they'll be able to find any holes in that system, just like the generations before them.
> For that matter, I'm surprised that HN automatically always accepts the "slippery slope" fallacy while lambasting it everywhere else.
Slippery slope arguments are not automatically a fallacy. They can be if the causative relationship is weak or if the slope is massively exaggerated. But if neither of these things are true, "slippery slopes" is just looking at the trends and expecting them to continue. You can't look at a linear graph and say "well, I think there's no most likely option from now on, it could go any way really" without an argument for why the trend would suddenly deviate. The internet had been tightening up and the walls have been closing in for a long time, why would that change?
In 2020, about 63 million Americans were parents to children under 18 [1]. That's about one in four adults. It's a small minority. Maybe 69% of all people have children in general, but most of those children are above the age of 18.
Not that it matters anyway. You picked out a single word out of my multi-paragraph comment with many arguments and ignored the rest. Even if parents of young children were the minority, does that mean they should dictate what the rest do for the sake of their safety? Like 85% of Americans are drivers, should the non-driving pedestrians start stocking up on bubble wrap? It's for the common good, after all.
And I keep thinking who can AFFORD to pay per token? I did a simple test - three small files and a prompt was nearly 10k tokens. Compared to my actual code base, where I use 5.2/sonnet to parse huge chunks of my code...I'd be burning hundreds of dollars per day if i was doing it per token rather than copilot - let alone the huge agent sessions where I use Opus and it has 50+ back and forward attempts.
Please note I do actually read every line of code these reckless hacks generate haha.
I've had single prompt to Opus consume as many as 13 premium messages. The Copilot harness is so gimped so they can abstract tokens from messages. Every person that started with Copilot that I know that tried CC were amazed at the power difference. Stepping out of a golf cart and into <your favorite fast car>.
It hasn't done that to me. It's worked according to their docs:
> Copilot Chat uses one premium request per user prompt, multiplied by the model's rate.
> Each prompt to Copilot CLI uses one premium request with the default model. For other models, this is multiplied by the model's rate.
> Copilot coding agent uses one premium request per session, multiplied by the model's rate. A session begins when you ask Copilot to create a pull request or make one or more changes to an existing pull request.
Sorry, I should have specified this was with GHC CLI. I suppose that might not behave similarly to the GUI extension. But it definitely happened on Thursday. One prompt, ctrl-c out and it said 13 premium messages used. It was reading a couple of large files and Opus doesn't seem to let the harness restrict it from reading entire files... just a couple hundred lines at a time.
and now I see your comment mentions that explicitly. The output was quite unambiguous. :shrug:
Hey! I'm a PM on the Copilot CLI team. This sounds like a bug, we should follow the same premium request scheme as the VSCode extension! If you still have the session logs kicking around, can you email them to me? It's my hn username @github.com
It seems like it's the cheapest way to access Claude Sonnet 4.5, but the model distribution is clearly throttled compared to Claude Sonnet 4.5 on claude.ai.
That being said, I don't know why anyone would want to pay for LLM access anywhere else.
ChatGPT and claude.ai (free) and GitHub Copilot Pro ($100/yr) seem to be the best combination to me at the moment.
Cat's out of the bag now, and it seems they'll probably patch it, but:
Use other flows under standard billing to do iterative planning, spec building, and resource loading for a substantive change set. EG, something 5k+ loc, 10+ file.
Then throw that spec document as your single prompt to the copilot per-request-billed agent. Include in the prompt a caveat that We are being billed per user request. Try to go as far as possible given the prompt. If you encounter difficult underspecified decision points, as far as possible, implement multiple options and indicate in the completion document where selections must be made by the user. Implement specified test structures, and run against your implementation until full passing.
Most of my major chunks of code are written this way, and I never manage to use up the 100 available prompts.
This is basically my workflow. Claude Code for short edits/repairs, VSCode for long generations from spec. Subagents can work for literally days, generation tens of thousands of lines of code with one prompt that costs 12 cents. There's even a summary of tokens used per session in Copilot CLI, telling me I've used hundreds of millions of tokens. You can calculate the eventual API value of that.
For $10 flat per request up to 128k tokens they’re losing money. 100 * 100k is 10m tokens. At current api pricing that’s $50 input tokens, not even accounting for output!
Having worked some time in huge businesses, I can assure that there are many corporate copilot subscribers that never use it, that's where they earn money.
In the past we had to buy an expensive license of some niche software, used by a small team, for a VP "in case he wanted to look".
Worse in many gov agencies, whenever they buy software, if it's relatively cheap, everyone gets it.
It might be a gym-type situation, where the average of all users just ends up being profitable. Of course it could be bait-and-switch to get people committed to their platform.
> I was a vegan for 10 years, and found out I had basically zero Omega 3 in my blood
I see your disclaimer, but just for more context, vegans can get Omega 3 without taking pills per se. Flax seeds are an excellent source. I often add a spoonful to a bowl of oatmeal or as a pancake topping along with fruit sauce and granola.
Most people can’t chew flaxseeds effectively, so they grind them first or swallow them whole. (They are tiny.) Nutrition experts do recommend grinding them first to release the fiber and the beneficial fatty acids. Flaxseeds are helpful for constipation and may lower cholesterol as well.
Ground flaxseed goes rancid easily, however, so it should be kept in the freezer until you are ready to use it. If you buy it ground, you wouldn’t have to use the blender or coffee grinder to break those seeds up before you have breakfast.
Flax seeds are a very tedious and inefficient way to get omega-3 as a vegan, particularly because they contain ALA, a short chain omega-3, which our bodies are extremely inefficient at turning into long chain fatty acids.
Funny - I feel the opposite about chia. Soaked and plumped is when I hate them. Dry on salads/etc. or just submerged in an active bowl I'm eating is when I like them most - the crunch adds texture to what I'm eating.
And better than taking pills for the former, add hemp hearts or flax seeds to your cereal. One serving of hemp hearts has 10 grams of protein and 12 grams of Omegas 3 and 6. Flax seeds are lower in protein but an even better source of Omega 3 in particular.
Never going to advocate against eating whole foods if they taste good! But beware, the ALA omega 3 fat in flax and plant sources is not the DHA and EPA omega 3 fats used by animal cells, and so it's not as potent as what's in fish.
The main problem with ALA is that to have the good effects attributed to omega-3s, it must be converted by a limited supply of enzymes into EPA and DHA. As a result, only a small fraction of it has omega-3's effects — 10%–15%, maybe less. The remaining 85%–90% gets burned up as energy or metabolized in other ways. So in terms of omega-3 "power," a tablespoon of flaxseed oil is worth about 700 milligrams (mg) of EPA and DHA. That's still more than the 300 mg of EPA and DHA in many 1-gram fish oil capsules, but far less than what the 7 grams listed on the label might imply.
Also, beware of omega 6 fats. Seed oils (corn, soy, canola) used in commercial food products are incredibly omega 6 dominant in terms of polyunsaturated fat content. Consequently, the ratio of omega 3 to omega 6 fats we consume has plummeted as food production has industrialized. Omega 3 fats are precursors to generally anti-inflammatory signaling compounds, whereas omega 6 fats are precursors to pro-inflammatory signaling compounds. The bias in fat intake leads to more pro-inflammatory signaling in the body, and a lot of alt health types have alleged this is a major causative factor in the obesity epidemic.
This is important for depression, because chronic brain inflammation as a cause of depression was one of the going hypotheses at least a decade ago when I last looked into all of this. Upping omega 3 intake is an intervention that can address chronic inflammation, which is potentially why it improves some cases of depression.
Pretty much nobody in the west needs more omega 6s these days. I hear even farmed salmon eat primarily corn and soy based feeds these days, meaning their fat ratio is skewed much more heavily toward omega 6 than wild salmon and fish.
I'm not an expert, but I've done a bunch of reading on this previously, and also skimmed the article which also mentions some parts of this.
First, when taking omega 3 supplements, you generally care about increasing the ratio of omega 3 to omega 6. Hemp hearts have much more omega 6 than omega 3, so they're not very effective for improving the ratio.
Second, hemp hearts contain ALA, while what you generally want to improve is EPA and DHA (this is also covered in TFA). The body can convert ALA to EPA and DHA, but it's not efficient.
So all in all, if Omega 3 for the article's stated benefits is what you want, this is not the way. I recommend looking into eating more fish, or if you want a vegan route, algae-based supplements. [0] is a decent source from the NIH about foods and their Omega 3 content, split by ALA/EPA/DHA.
The ratio of Omega 6 to 3 needs to be below 4:1 to be a good source of Omega 3, and hemp hearts are at 3:1, so they're listed as a good source of Omega 3.
Flax seeds are even better just for Omega 3 at 1:3, but hemp hearts have other benefits, like more protein, which is why I called them out. That said, I eat a fair amount of flax seeds as well.
Just to reiterate, both of those (hemp hearts and flaxseed) only contain ALA, while what you're generally looking for is EPA and DHA. TFA also explicitly mentions it's only talking about EPA.
This is not to say that they're unhealthy of course.
EDIT: see the sibling comment by code_biologist, it's much more comprehensive than what I've written.
try the tools. Really. If you are remotely interested in tech or AI, try the tools
Copilot this is not. You may be trolling of course. There are huge steps between these various tools, if you try them, for a smidge of investment, it will become obvious what the trajectory is.
It is like saying "I don't handwrite anything, I care too much about line spacing, I only use a dot matrix printer" when some one is trying to sell you a calligraphy pen and coloured inks, and you have only tried a ballpoint pen. You might be the wrong market, but they are not even close in use case and application.
I'm not trolling. I'm just not aware of major differences between them.
When I make a change with a Copilot Agent, it checks for issues, builds my project, runs tests, and iterates until things work. Multiple agents can do that in parallel.
My impression was that this does more or less the same thing.
That said, I'm definitely open to learning more about them both.
What are the advantages of this in your experience?
It is worth an install; it works very differently than an agent in a single loop.
Beads formalizes building a DAG for a given workload. This has a bunch of implications, but one is that you can specify larger workloads and the agents won’t get stuck or confused. At some level gas town is a bunch of scaffolding around the benefits of beads; an orchestrator that is native to dealing with beads opens up many more benefits than one that isn’t custom coded for it.
Think of a human needing to be interacted with as a ‘fault’ in an agentic coding system — a copilot agent might be at 0.5 9s or so - 50% of tasks can complete without intervention, given a certain set of tasks. All the gas town scaffolding is trying to increase the number of 9s, and the size of the task that can be given.
My take - Gas town (as an architecture) certainly has more nines in it than a single agent; the rest is just a lot of fun experimentation.
Yes he is on an extended manic episode right now - we can only sit back and enjoy the fruits of his extreme labor. I expect the dust will settle at some point, and I think he’s right that he’s on to some quality architecture.
In your post history you say you have never programmed. Why are you so sure it produces code of value?
This is so prohibitively expensive in its wastefulness that blithely telling strangers to try the tools likely means you either haven't tried it, or have money to burn.
Your cold-outreach emails will be flagged as spam by some people, even if they comply with CAN-SPAM. And you should comply with it by having a clear opt-out mechanism and your physical address in the footer, but some people will flag your emails regardless.
I've had this happen in my startup. It hurt the deliverability of our transactional emails (notifications etc.). I'd go with a separate domain if I had to do it over.
reply