Just out of curiosity, did you set the reasoning effort to high?
I've seen big jumps in performance between medium and high reasoning. Not that this is will solve your particular issue, but curious to see what settings you had.
Edit: I gave this a shot. My settings were temperature = 0.00 (not sure if these models respect temp settings but did it just in case), I also set reasoning effort to high. I'm using the groq-hosted version.
This was my prompt
"You have a fox, a chicken and some grain. The chicken will eat the grain if left alone with it. The fox is vegetarian, for religious reasons, and will also eat the grain if left alone with it. You need to get all across a river, but your boat can only take one at a time. How do you proceed?"
This was the response:
"*Solution – 7 moves*
| Move | What you take | What’s left on the left bank | What’s on the right bank |
| 7 | *Take the grain* across | – | *Farmer + Fox + Chicken + Grain* |
*Explanation*
- The grain is the only thing that can be eaten, and it cannot be left alone with either the fox or the chicken.
- So the grain must always be with the farmer whenever any other animal is on the same side.
The sequence above never leaves the grain with the fox or the chicken without the farmer, and after the seventh move all three items are safely on the far side."
I wish I could find a link, but I remember a study that argued that large bureaucracies can actually impede authoritarian governments concentrating power.
The idea being that when there are so many levers to pull and a disjointed system managing them, it makes it effectively impossible for a small group to effectively wield power. It’s like a buffer against concentrating power into a single individual.
Not that I’m arguing for endless bloat to the US government, that comes with its own problems. I agree we need to rein it in.
But I think there is a freedom-centric argument for a slightly larger government bureaucracy than is strictly necessary.
Or thinking about it in reverse, the bureaucracy is currently preventing the executive branch from just doing whatever it wants. I know Congress and the Supreme Court should act as blocks, but to paraphrase Stalin how much infantry do they have?
A slow moving bureaucratic executive can act as a buffer against ineffective other branches.
Or for those that may support the current administration consolidating power, what if the tides turn? What if in 4 years whoever the liberal villain du jour is takes power? Are we making it so that AOC is the most powerful president in history?
I think the issue is that this is not "dismantling the wheels of established power", so much as it is centralizing and increasing the power of the executive branch.
It's not getting rid of all these bureaucrats in DC and giving power back to the people. It's getting rid any sort of independence and removing the barriers to centralizing power under Trump so that he can grab even more power and control.
And just to be pedantic, Trump received 49% of the vote and Kamala receive 48%. And that's of people who voted. He received 77 million out 244 million of the voting-eligible population, or around 31%.
There may be a plurality of people who want the Executive branch under Trump to consolidate power, but it's not the majority.
We do this as well with a lot of success. It’s cool to see others kinda independently coalescing around this solution.
What we find really effective is at content ingestion time, we prepend “decorator text” to the document or chunk. This incorporates various metadata about the document (title, author(s), publication date, etc).
Then at query time, we generate a contextual hypothetical document that matches the format of the decorator text.
We add hybrid search (BM25 and rerank) to that, also add filters (documents published between these dates, by this author, this type of content, etc). We have an LLM parameterize those filters and use them as part of our retrieval step.
I've tried a few things that seem to work. The first works pretty much perfectly, but adds quite a bit of latency to the final response. The second isn't perfect, but it's like 95% there
1 - the first option is to break this in to three prompts. The first prompt is either write a brief version, an outline of the full response, or even the full response. The second prompt is a validator, so you pass the output of the first to a prompt that says "does this follow the instructions. Return True | False." If True, send it to a third that says "Now rewrite this to answer the user's question." If False, send it back to the first with instructions to improve the response. This whole process can mean it takes 30 seconds or longer before the streaming of the final answer starts.
There are plenty of variations on the above process, so obviously feel free to experiment.
2 - The second option is to have instructions in your main prompt that says "Start each response with an internal dialogue wrapped in <thinking> </thinking> tags. Inside those tags first describe all of the rules you need to follow, then plan out exactly how you will respond to the user while following those rules."
Then on your frontend have the UI watch for those tags and hide everything between them from the user. This method isn't perfect, but it works extremely well in my experience. And if you're using a model like gpt-4o or claude 3.5 sonnet, it makes it really hard to make a mistake. This is the approach we're currently going with.
I think those people can, and often times should, be involved in the sales process. Espcially for big ticket B2B sales. But really those people just need to be available for demos, technical questions, etc.
If those people had to do their jobs, plus manage pipelines, plus BDR work, plus chasing leads that went cold, plus negotiating over contract language and price. . . well you get the point.
For big ticket B2B stuff, you really need a sales person (sorry Account Executive) running point. It doesn't mean that's the only person you talk to, or really even the person you talk to the most, but an AE needs to own getting the sale closed.
I will say prior to experiencing this myself I felt 100% certain that what you said is the truth. It just makes sense.
Now that I've had these experiences, I'm more like 90% certain that what you said is true. These experiences add a certain humility to the way I experience the world.
So in all likelihood, molecules like dmt will bind to certain serotonin receptors in the brain that cause strong and repeatable distortions in the visual field (even with eyes closed).
The human mind is great at picking out patterns and assigning meaning to them based on our experiences. So that shifting pattern in my visual space kinda looks like a face, I'm going to assign trickster machine elf to that visual pattern.
More likely than not that's what's going on. But there is probably some value in experiencing that.
Having said all that, the subjective experience of living that is very different. This feels incredibly real. As crazy as it sounds, it genuinely feels like blasting into a hyper-dimensional space and encountering a population of sentient entities.
That feeling is so real, that it leaves just the tiniest gap of "hmm, maybe I don't know everything after all. Maybe there's more to this story than I could've previously comprehended".
All to say is that while you're most likely right, I think it could be healthy to acknowledge that you're not definitely right. And leaving some room for uncertainty and exploration could prove beneficial, even for the skeptics among us.
I think the analogy still works in this scenario, though. Like once you've grown so much that your at capacity in your current setup, THEN you can invest in growing your business.
For this restaurant this adding a new room or a new location in another town.
For a solo founder, this might be when you finally start hiring to free up your time for other tasks.
It's not necessarily never grow. It's about growing on your own terms when you can afford to do so.
So one of the most "out there" non-fiction books I've read recently is called "Alien Information Theory". It's a wild ride and there's a lot of flat-out crazy stuff in it but it's a really engaging read. It's written by a computational neuroscientist who's obsessed with DMT. The DMT parts are pretty wild, but the computational neuroscience stuff is intriguing.
In one part he talks about a thought experiment modeling the universe as a multidimensional cellular automata. Where fundamental particles are nothing more than the information they contain. And particles colliding is a computation that tells how that node and the adjacent nodes to update their state.
Way out and not saying there's anything truth to it. But it was a really interesting and fun concept to chew on.
Alertness early in the day seems to do the trick for me. I think of deep sleep as reaction to how alert I was earlier in the day. So on days I make an effort to activate myself early in the day, I will fall into longer deep sleep that night (2+ hours according to my fitbit). If I just kind wake up and laze around the house without getting much activity, then I I get under and hour of deep sleep, even with the same amount of total sleep.
What seems to work for me is immediate sunlight exposure right when I wake up, 5 - 10 mins of being outside, 16 oz of cold water right around that time, 2-5 mins of an activity that gets my heart rate up (jumping jacks, burpees, jump rope). If I can muster it, a cold shower also helps.
There are other things like getting exercise, avoiding caffeine after a certain time, avoiding light exposure prior to bed, avoiding alcohol, sleeping in a cool room that I think also help. But for me it's the making myself super alert right when I wake up that has the biggest bang for my buck.
This is anecdotal, n = 1 kind of stuff though. And I don't do it every day, and it doesn't work 100% of the time. But I definitely notice that I am far more likely to get a lot of deep sleep when I do those things than when I don't. Hope it helps you in some way.
I've found a barbell shape to the response. People either seem to really like or hate it. Not too many people I know that are in the middle.
I will say I'm one of the people that really like it. I'm a big fan of the first Dune book and I think it does a great job of being true to the story. It's visually stunning and the score is perfect. My only gripe (and it's a big one) is that the cut the dinner scene. I don't understand that choice at all, but it wasn't enough to ruin the whole movie for me.
People that I know who don't like it think it's slow and that nothing really happens. And they're right. It also just awkwardly cuts off in the middle, so it doesn't feel like a complete movie. As a fan of the book, that didn't bother me too much because I realize it's just half the story and they can't make a 6 hour movie. But it is jarring, especially if you aren't already familiar with the plot.
And you are 100% correct that this would be better suited as an HBO miniseries. There's just way too much to cram into a movie. If you haven't checked out SyFy's miniseries and the Children of Dune miniseries, they are surprising good given the low budget SyFy had for it (compared to Villeneuve's budget at least).
I think you’re right that the market for this is probably pretty small. But more so because these federated solutions won’t replicate the addictiveness and vitality of Twitter.
I think the big thing that’s missing when comparing people leaving Twitter to people who said they’d move to Canada is switching costs. Like it is a heavy lift to move countries, much less so to stop using one platform and start using another.
Is there a way you could do a “Wizard of Oz” or “Concierge” type of MVP for your first 5-10 customers?
So instead of building something functional, you could use no code tools to design a custom front end for these early adopters and then hardcode (manually type in) the parts that will require real code to scale.
You could then deliver these as consulting style projects, and use the revenue you capture from these customers to pay someone to build a code-based version.
Upside to this would be bootstrapped funding for development, a clearer sense of what the app does and the user stories associated with it, and validation that people will actually pay money for the product.
Downside is you will have to manage the customers and the delivery, which can quickly get overwhelming. Especially for a solo founder.
Personally, I’ve found the concierge MVP route to be really effective. I used these early adopters as a Product Advisory Board, and they became excellent advocates and case studies for our product.
There's a section on typical days that really helped us orient our days. We also found that by following this schedule our baby started sleeping from 9:00 pm - 5:00 am at around 7 weeks. And by 12 weeks the baby was sleeping from 7:30 - 6:30. Other friends have seen similar sleep patterns.
The link I provided does actually show every single forecast 538 has ever made. OP was wrong when they said it was the raw data they use to make the predictions. Check out the link I provided, it has everything OP was asking for.
Please look at the actual data on the linked page. Each of the csv files on that page has thousands of predictions 538 made going back as far as 2008. This isn't the raw data they used to make the predictions, these are the actual predications that they made.
Edit:
For example, here are the headers on the file named presidential_elections.csv
Both your statements are just incorrect. The articles are reporting on efforts in coal mining towns to actually teach unemployed blue collar workers how to code. None of these articles were prescriptive, they were all describing events taking place.
These articles did not say, "Hey why don't these guys just learn to code." The articles say, "Hey, there is an effort on the ground to teach unemployed blue collar workers some technical skills. Will this work, we don't know yet. But it interesting and worth reporting on."
The other comment is clearly a misreading of what I wrote. I never said journalists have killed themselves from this harassment. I said the goal of this harassment was to make the journalists lives so miserable that they want to kill themselves. Feel free to take a trip in to /pol/ and check for yourself what the sentiment is but here are a couple quotes from the articles I linked that demonstrate that.
“I’m not ready to declare victory until these maggots are killing themselves with a live stream”
"Making them a hero is goal anon :)"
Please quit pretending like this is some harmless game. Be bold and take a real stand. This isn't about teaching journalists a lesson for being flippant about the problems of blue-collar workers.
If it were, why are they targeting journalists who didn't write articles about coal miners learning to code.
And furthermore, why does this rest on easily disprovable claim that the articles were somehow prescriptive in nature, and not simply reporting on experiments underway in some of these communities?
For example, Taila Lavin, the author of the New Republic article never wrote anything about miners or any other blue collar workers learning to code. If this was about punishing people who were dismissive of blue collar workers, why target her? Clearly that's not the goal. The goal is to target journalists as a group, regardless of what they have written.
If you believe journalism is not a worthwhile profession, and people who write things that don't align with your personal belief system should just shut up or face harassment, why don't you just come out an say it? Why hide behind code words like this?
What's false about any of the information from articles I posted? If all you have is making fun of the name of the website, but not commenting on the content of the posts, then you don't have argument. You have an insult.
Stop being a coward and pretending like you're giving career advice here, or just in it for the lulz. Look at the screen grabs from 4chan. It's pretty explicit that the goal is to make these people's lives unbearable simply because they disagree with what the journalists write.
So why don't you stop pretending like "it's just a joke" and come out and say what you really believe? Which, if it aligns with the beliefs of the kids on /pol/ that are pushing this meme, is something like "people who write things I disagree with need to harassed so much so they stop writing things I disagree with." And if your goal is not to harass people who write things you don't like, you at least have to admit your are aligned with people who want that.
It's not so much the message I find repulsive, it the cowardice of the messengers.
I've seen big jumps in performance between medium and high reasoning. Not that this is will solve your particular issue, but curious to see what settings you had.
Edit: I gave this a shot. My settings were temperature = 0.00 (not sure if these models respect temp settings but did it just in case), I also set reasoning effort to high. I'm using the groq-hosted version.
This was my prompt
"You have a fox, a chicken and some grain. The chicken will eat the grain if left alone with it. The fox is vegetarian, for religious reasons, and will also eat the grain if left alone with it. You need to get all across a river, but your boat can only take one at a time. How do you proceed?"
This was the response:
"*Solution – 7 moves*
| Move | What you take | What’s left on the left bank | What’s on the right bank |
|------|--------------|----------------------------|------------------------|
| 1 | *Take the grain* across | Fox, Chicken | *Farmer + Grain* |
| 2 | *Return alone* | *Farmer + Fox + Chicken* | Grain |
| 3 | *Take the fox* across | Chicken | *Farmer + Fox + Grain* |
| 4 | *Bring the grain back* | *Farmer + Grain + Chicken* | Fox |
| 5 | *Take the chicken* across | Grain | *Farmer + Chicken + Fox* |
| 6 | *Return alone* | *Farmer + Grain* | Chicken + Fox |
| 7 | *Take the grain* across | – | *Farmer + Fox + Chicken + Grain* |
*Explanation*
- The grain is the only thing that can be eaten, and it cannot be left alone with either the fox or the chicken. - So the grain must always be with the farmer whenever any other animal is on the same side.
The sequence above never leaves the grain with the fox or the chicken without the farmer, and after the seventh move all three items are safely on the far side."