This means that, in practice, it functions more like an agent than a streaming text model. That in turn breaks trajectory analysis which not only makes distillation harder, it also makes it much harder to improve your own harnesses and agents built on top of the model.
HN user
lukemercado
Foam | Staff Founding Eng | In-Person, San Francisco | Full-time (M-F 9am-7pm) | $300-$400k/yr + equity
Foam is building a Better Sentry on an observability stack purpose-built for AI. We automatically identify and root-cause breakages across systems with zero prompting and zero configuration then DM the engineer responsible, not the git-blamed engineer.
We're looking for a hands on keyboard Staff Founding Engineer who's been unshackled by modern coding agents.
Here's a few of the problems we'd love your help on:
1. Reducing our onboarding time from hours to seconds
2. Scaling our telemetry engine into the petabytes
3. Collapsing our latency from minutes to seconds
4. Beating our intern at chess (ELO 2200)
You'll be joining an incredible team:
1. Ben, Software Engineer Intern, https://www.linkedin.com/in/ben-hung/, beat out dozens of seniors on our coding challenges while holding a 3.9+ GPA in Berkeley EECS and 2200 in Rapid chess
2. Nyx, Founding Research Engineer, https://www.linkedin.com/in/nyx-iskandar/, Prev VP ML@Berkeley, Open AI Launchpad Researcher, Berkeley EECS (2.5yrs w/3.98 GPA)
3. Luke, Co-Founder & COO, https://www.linkedin.com/in/lukemercado/, Fmr CTO & Co-Founder @ Daybreak Health (S20, Series B), self taught SWE (2013), college reject
4. Shawn, Co-Founder & Engineer, https://www.linkedin.com/in/shawn-krisman/, early @ Yelp, Everlane, Reddit, Patreon & Affirm
5. Perla, Co-Founder & CEO, https://www.linkedin.com/in/perla-gamez/, delivered ~$X00M revenue @ Affirm as a SWE, Berkeley CS ('18), Honduras National Karate Champ
We've raised $10M in seed funding from Khosla Ventures, Max Levchin, The House Fund, and South Park Commons and others.
Visas Supported: H1B Transfer, H1B1, TN, E-3, OPT-STEM
Email careers at foam dot ai with your resume and we'll get back to you within 24 hours.
That was lovely, thank you for sharing.
It seems the core of the game is correctly understanding the victory function (time to complete refinement of all resources) and then applying bottleneck theory to correctly optimize each stage.
Definitely a joyous little ten minute romp.
foam.ai | Founding Engineer | $180-$300k | 0.25%-2% | SF (onsite) | Full Time
foam is building a new observability stack that captures, traces, and links everything, purpose-built for AI. We use this foundation to feed complete context into AI to automatically identify and root-cause breakages across entire systems.
* We’re a five person team. Perla (CEO) was the 5th engineer in the consumer facing team at Affirm (pre-IPO).
* We've raised $9.9M in seed funding from Khosla Ventures, Max Levchin, The House Fund, and South Park Commons
* We’ve got active betas at Orb, Braintrust, Plaid & TogetherAI.
We’re looking for engineers who want to take point and work hand in hand with Perla on one or more of our core systems:
1. The Telemetry Engine - Think of this as a major extension of OTEL custom built to capture signals that have historically been inaccessible to humans.
2. The Root Cause Analysis Engine (aka Solver) - This is currently an agent swarm, but evolves every week in search of better and better root cause analysis.
For a bit more on the role, check out: https://foam.ai/careers
If putting an end to awful on-calls and building a system that can deliver self-healing software is exciting to you, send an email to luke -at- foam dot ai with your resume and mention you came from HN.
foam.ai | Full Stack Founding Engineer | SF (onsite) | Full Time
At foam we’re building the first telemetry and production system designed for LLMs, giving them complete visibility into and control over running software. Today we’re making money by fixing bugs as they happen, tomorrow we’re the data layer for all production observability tooling.
* We’re a six person team. Perla (CEO) was the 5th engineer in the consumer facing team at Affirm (pre-IPO). Shawn Krisman (CTO) was an early engineer at Yelp, Reddit, Patreon and architect at Affirm (pre-IPO).
* We've raised $8.4M in seed funding from Khosla Ventures, Max Levchin, The House Fund, and South Park Commons
* We’ve got active betas at Perplexity, Orb, Braintrust, Plaid & TogetherAI.
If being #7 is exciting; reach out: send an email to luke -at- foam dot ai with your resume and mention you came from HN
Farm animals grown with their brains shut off, used as compute substrate for biological neural networks, while their biological functions are controlled remotely.
I’m sorry, you were working on what? Where does one learn more about this concept?
From first principles I'd expect the functional limit not to be one of power generation but of energy transfer. The limit probably manifests as wiring and motors overheating in a full electric drive (I don't think any subs are...) or as sound, heat and fatigue in reduction gears and gear shafts.
Location: San Francisco, CA, USA
Remote: Yes, but in-person preferred
Willing to relocate: No
Technologies: Python, Ruby, React, Salesforce, TypeScript
Résumé/CV: https://docs.google.com/document/d/1ux2NKFfinMR_n5d41uge0O2FstDSc-oI/edit?usp=sharing&ouid=110631342450297019344&rtpof=true&sd=true
Email: luke+hn@lukemercado.com
I'd be particularly interested about head of eng roles at startups or director of eng at a scale up. After the co-founding experience I'm also curious about moving into product management or sales in some capacity. For the right company I'd do some software engineering.Location: San Francisco, CA, USA
Remote: Yes, 3+ years of remote experience, also interested about going to an office.
Willing to relocate: Yes
Résumé/CV: https://docs.google.com/document/d/1TggpfOgWs7R30Ixyf4O1zwlS...
Email: luke at lukemercado dot com
About Me: Former CTO of a YC backed, Series B company, looking for new opportunities.
LinkedIn: https://www.linkedin.com/in/lukemercado/
Did the 350 california sale go through? I thought it was just an offer?
Just googled it, and shit, it did: https://www.sfgate.com/local/article/350-california-sf-offic... Holy hell. Now to find out if they land bank it or blow it up.
It takes a lot of exercise to burn off excess calories, and virtually no effort to simply not consume them.
and virtually no effort to simply not consume them.
I don't think this is true. The executive function required to manage the dopamine cycle that food generates is incredibly taxing. I honestly think this is a result of our food systems being incredibly optimized to drive purchasing decisions...
Separately, this reminds me that I should eat some veggies.
I don't see how wealth is lost when it is just being redistributed. The workers didn't disappeared, they are just living elsewhere, bringing business opportunities and improving across many different places.
If I'm recalling the theory of it correctly, the whole reason Business Clusters exist (and form) is because there is, essentially, value in being together. Or at least, there's economic value in reducing the logistics required to move goods from one part of the value chain to another. For this reason I don't suspect we'll see the dispersal of manufacturing business clusters.
A huge part of my digging was trying to understand whether this applied to Silicon Valley or not. I decided that it did, but not because of the software, instead it was because of the VC + Startup relationship. VC's want to meet the people they're giving millions of dollars to and startup founders want to meet the people they're giving control (or massive influence) of their company to. It remains to be seen whether Zoom satisfies this need or not.
Perhaps it will be healthier for the economy as a whole than having only a handful of cities as beacons of investment.
It might. Technology has changed dramatically since we last had material distribution into the suburb and rural areas of the nation. Maybe that will overcome the benefits of centralization.
In general, I'd observe that there's been much more of a reset to pre-pandemic norms than many anticipated.
I haven't traveled well enough recently to agree or disagree, but this matches my reading, so I'll take your argument.
But SF could well end up being something of an anomaly--which would be especially bad news for SF given it may not be a broad-based crisis that the government at the national (or maybe even state) level is going to be especially concerned about.
Yeah... That's why I think we're somewhere between Theories 3 and 4. I was really hoping that the city supervisors would see that and start shifting SF's office core to be a more desirable and exciting place by reducing the barriers to entry for new brick and mortar businesses. Think more streets shut to cars, more bike only roads, massively more efficient and cheaper stall and cart permitting along with expanded areas of operation. Roll this together and you could essentially crawl the existing temporary street markets around the Ferry Building into the urban core. This would also provide an avenue for businesses to start and grow to fill the empty retail in the buildings around them, much the same that food trucks have become a stepping stone to brick and mortar restaurants.
Disclosure: I live in SF and co-founded www.daybreakhealth.com which is fully remote. I am, 100%, part of why this tower, and others like it, are empty.
"[...] Mayor Breed, who in an interview earlier this year said that “for this city to be thriving, we need people back in the office.”"
This is such a frustrating statement. If I were going to make this point I would say "for this city to be thriving in its current incarnation, we need people back in the office." Sadly, the mayor, as espoused above, and at least one city supervisor (Matt Dorsey¹), simply do not seem open to the rebirth necessary to make this city thrive again.
In 2020, in preparation for the debates with my co-founders on whether we should establish a fully-remote or fully-in-office culture I spent a lot of time trying to understand what remote work would mean for the Silicon Valley Business Cluster² and San Francisco in particular. I built a working theory for how Silicon Valley functions as a business cluster, San Francisco's role within that cluster, and San Francisco's city finances. Based on that working theory I built a cone-of-possibility broken out into four main scenarios that looked something like this:
* Theory 1 - Remote Work is a Pandemic Only adventure. Workers will exit the major cities, first in fear, then in pursuit of better personal finances. The pandemic will end, and the functions that drew workers into the city in the first place will reassert themselves. Workers will flock back to business clusters, the cities within them, and the offices in which they used to work.
* Theory 2 - Remote Work isn't sustainable for workers in most industries. Workers will exit the major cities, first in fear, then begin to hate their new normals. They will return to the city, and to the office, possibly before the pandemic ends.
* Theory 3 - Remote Work in Software is here to stay. Business Clusters that relied less heavily on Software workers, such as LA's film cluster, or the Boston / Cambridge BioPharma cluster would see a return to the office. Yes, they would have a lower over all demand for office space, but we'd be looking at a mid single digits³ reduction. Meanwhile San Francisco's office buildings would become a ghost town. This reduction in worker need for office space in SF would drive a cascade collapse in its office districts. First the streets are empty of pedestrians, which collapses all the local retail, which draws in more of the unhoused, which trash the place due to lack of supporting infrastructure, which drives up the cost of keeping the streets from looking like the Tenderloin. All the while, commercial leases lapse or go up for sublet in alarming volumes. These forces of reduced software workers, reduced retail spend, and reduced office rents combined to create a financial collapse in SF's city budget, anywhere from 10 to 20% of city revenues. In the worst case, with vacancy reaching into the 70 or 80% range, SF's budget craters by as much as 30% forcing a material reduction in city services which exacerbates the unhoused problem, creating a vicious cycle.⁴
* Theory 4 - Remote Work is here to stay for all knowledge work. In this scenario, all of the financial woes that play out for SF in Theory 3, play out for any city that has a major knowledge worker dependency. The breadth of the commercial real estate collapse begins to threaten the solvency of many banks. The workers, faced with the isolation of in-home work begin to branch out socially in their evenings and weekends. We see a resurgence in hobbies that bring people together.
By my understanding we're currently somewhere between Theory 3 and Theory 4. I think my theories were a bit aggressive on the potential collapse scenarios, but I'm seeing enough of what I predicted to feel comfortable in my reasoning. With that in mind, let's talk about how this ends: The lack of demand for office buildings in many cities results in the collapse of the system used to finance them. Banks and the hedge funds, REITs, pensions and other debtors take a monstrous haircut as the buildings are sold at huge (60%+) losses. A few will make it out OK, having held the buildings long enough to have turned a profit, but none of them walk out of this with their financial projections intact. The buyers of these buildings are, predominantly, in it to blow them up. The buildings are torn down and replaced with purpose-built housing and specialty offices (think biopharma or cultured-meat infrastructure).
This article, and 350 California Street in particular, represent the pin hole in the proverbial dam. I wait with baited breadth.
¹. I spoke with Dorsey for several hours during his District 6 Supervisor election campaign on this topic as I was living in his district. His views can be concretely summed up as a deep belief that tax policy alone can get office workers back. He seems unable, or unwilling, to distinguish between office tenants and office workers.
². https://en.wikipedia.org/wiki/Business_cluster#The_Silicon_V...
³. This is purely a guess, I base this on nothing but my assumption of how many software folks work in these industries.
⁴. That's leaving out the potential residential real estate cascades which in Silicon Valley could actually kill the business cluster itself.
I really like Tyme for time tracking. It works quite well for me, is dirt cheap, has a great UX across both my iphone and my macbook and gives me flexibility around start and end times so it's easy to keep it roughly accurate.
That said, I don't use it for detailed task observation. For example while writing this my "management" timer is running and this is definitely not managing my team. I use it for directionality more than anything else.
Short, but interesting little piece. I specifically enjoy how it's not just a rant about how the Player / Coach concept is bad. It's also an explanation for how to make the Player / Coach concept work well.
Saving face and preserving "decision making credibility" means that taking advice is simply not on the cards for a large portion of the middle management class.
While I'm happy to admit that this construct may be true in practice; it is _deeply_ infuriating that so many people's calculus nets out in this manner. It's infuriating to me, primarily, because I simply don't understand. By my understanding, "decision making credibility" comes _exclusively_ from *being right*. If you're optimizing for this metric, then how you get there should be an almost irrelevant footnote.
Yet here we are; with a non-trivial percentage of managers coming to the conclusion that the correct answer is to not take advice.
I've been exploring this concept a bit at my startup for our B2B offerings. Where I've landed is a robust set of feature toggles that Sales and Account Management configure. Within the confines of your Microsoft Word toolbars analogy; since the software we offer is similar to SaaS and delivered via a web browser this effectively means we can "hide" the toolbars that the client isn't interested in.
This is all pretty early days, but my hope is that we can iterate the core "have feature toggles" concept to the point where we have the ability to turn functionality and function accessibility on across many verticals. I hope that one day internally we can configure things a per-customer and per-user-persona basis. I also hope that one day we can expose some of those switches to admins and/or everyday users, possibly with paywalls or other strictures.
We'll see where it all lands :D
This whole article is a beauty.
FTX worked fine: People liked its technology, and it seems to have made money. The problem was in its balance sheet, which was full of snakes, and its governance, which put all the snakes there.
Fucking lol.
Everything you've said is logical, but the radius you are describing is multifaceted.
No disagreement there.
There are regulations, inertia, cultural and ideological factors, and all sorts of other considerations that come into the capital vs. labor struggle and recreate the radius even with remote technology.
No disagreement there either.
In our current world, American salaries are inflated for macro reasons, but the shift you are describing may very well flatten the skilled salaries for currently wealthier regions and increase those for poorer regions.
Seems we agree. A pleasure chatting with you.
You are paid based on how difficult you are to replace. Hence the location based pay
You're partially right. Pay is very much about how difficult you are to replace. However Location in many industries is no longer the dominant input.
Location is the dominant input in an office-centric model, but isn't in a Remote model. In an office centric model, the company would have to pick where to locate the office and that location would have a "commute radius" in which they could effectively expect their employees to willingly come in to the office from. This meant that any physical office's "commute radius" was effectively its talent pool.
So if you put your office in a high cost of living area, you now have to pay enough for people to live in that area or be willing to commute in from lower cost of living areas. This reality is the foundation for the "location based pay" detente between Labor and Capital.
It's worth noting that these things are also interdependent and self reinforcing. Eg: An area (like Silicon Valley, or any other business cluster) has a higher density of talent, which is more attractive to employers, which means more employers want to be in that area. More employers means more competition for that talent, which drives wages up. As those wages then filter into the market you see the price for supply constrained necessities (houses typically) go up, and often times that will drive cost of living up.
In a remote work world, this entire cycle is dramatically disbursed and the core foundation of the "location based pay" detente between Labor and Capital is shattered. What this ultimately means is that Capital no longer pays for Talent + Location + Competitive wages within that location. Instead they pay for Talent + Availability (timezones mostly, but internet connectivity is in there as well). This should, in theory, drive a larger talent pool, which, depending on demand for that talent, might lower wages in a given "Availability". If it does, then the cycle starts anew as Employers start to seek out ways to be in that "Availability".
The big question one needs to ask is whether there is enough demand to suck up all the talent within the workable "Availabilities". If there is, then wages stay the same or go up. If there isn't, then wages go down.
My personal opinion is that any company doing "location based pay" in a remote work world is shooting themselves in the foot (unless they have market power in those locations).
Would you happen to know of an example of "tag-along rights" that I could bring up at my next board meeting? I'm curious what the implications of such a thing would be and having sample text would be incredibly valuable.
I'm sorry, I can't resist. What is a "fuzzy pumper" in this context?
This seems like an EXCELLENT way for junior engineers to finally get out of the "no job can't get a job" box. Super curious to see what comes of it, and to recruit those who complete the fellowships.
Daybreak Health (YC) | Lead Salesforce Engineer | USA Only (legal reasons) | Fully Remote (and staying that way) | Full-Time
I’m trying to hire for a Lead Salesforce Engineer and I am just utterly failing. I know I’m looking for a unicorn, but I’ve gone looking for unicorns in other roles (frontend, backend, data) and done well enough. How do I hire a kick ass salesforce developer?
I need them to do the following:
* 50% working with our Operations team to optimize and extend existing processes within Salesforce
* 50% working with our Engineering team to couple Salesforce to our existing Application (think making salesforce call our APIs at the right times)
The rub here is that I need someone who is technical enough to drive the salesforce components in the right way (maintainable, well structured, etc), has the devops chops (within the salesforce ecosystem) to stand up a scalable and workable SDLC for it and has the administrator experience to define the permission sets and profiles required to keep the Operations team moving but not breaking things as everything else happens.
Here’s the exact job description: https://daybreak-health.breezy.hr/p/70b654404f64-lead-salesf... (we could flex the salary a bit for the right candidate)
If you want a taste of what Daybreak's Engineering org is like, check out our Technical Vision -> https://daybreakhealth.notion.site/Technical-Vision-2021-09-...
Uh, yeah. My business budgets out 20k/year for insurance premiums on the employer side per employee. This might be an old shorthand, but it's what we use in our models. Some employees pay extra out of pocket for added dependents.
Yes, at first that is SUPER important, but wouldn't you ultimately want to fully decouple it from employers and just go to income taxes per individual for it?
The distinction is subtle, but in my mind having it structured as an individual paid income tax would yield better results for edge cases and not incentivize bad corporate behavior like running 1099 contractors instead of FTEs.
Daybreak Health | Senior Ruby Engs and Lead Salesforce Engineer and Lead Data Engineer | REMOTE (USA ONLY) | Full Time | https://daybreak-health.breezy.hr/?&department=Engineering#p...
Our mission is to create a world where every young person benefits from mental health support. We’re a YC-backed company that raised its Seed from Maven Ventures and just raised our Series A from a Top 10 VC (announcement coming soon). We’re a rocketship, with our revenue more than tripling between July and November of 2021, and more acceleration in sight.
Take a look at our JD's, each open role has a bespoke onboarding plan and will result in a role where you've got substantial autonomy and serious input into the business. If you'd like to enquire a bit more softly, feel free to reach out to me at luke at daybreak health dot com
I think for politicians above a certain level, we should go full Rawls. When you leave office, a random citizen from your constituency is selected, your net worth is set to their current net worth, and for a period of k years after leaving office, your income is funded by the public to match the income of that constituent.
This is a novel idea to me. Are you referring to https://plato.stanford.edu/entries/rawls/? If so, could you elaborate a bit on the philosophical underpinnings of this idea? My meager googling did not raise anything substantive.
Almost all combat is N+1. The holy grail is asymmetric weaponry/tactics. Unfortunately all asymmetric weaponry/tactics is eventually countered through R&D. The critical question is whether one combatant can gain a strategic edge over another during the novelty period of the asymmetric weapon/tactic.
Ah, the joy of the Military Industrial Complex's feature treadmill.